JP5348754B2 - Search support apparatus and method, and computer program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To retrieve and present a person highly relevant to a search object by using e-mail. <P>SOLUTION: This search support device 1 generates a new thread ID when a message ID of another related e-mail acquired from header information of the e-mail is not registered in a mail table, and acquires a thread ID from the mail table when the message ID is registered. A record with the message ID acquired from the header information, a sender and a recipient, and the thread ID set thereto is added to the mail table, and a keyword obtained by morphologically analyzing the subject of the e-mail is associated with the thread ID and written in a thread table. Upon receiving a search condition, the thread ID of a corresponding keyword is acquired from a keyword table, users set as a sender or a recipient and the number of settings are acquired from the record of the mail table specified by the thread ID, and a user having the larger number of settings is output by thread ID. <P>COPYRIGHT: (C)2010,JPO&amp;INPIT

Description

  The present invention relates to a search support apparatus and method, and a computer program.

In the information-oriented society in recent years, it has been difficult to find desired information from a huge amount of accumulated information. In an organization such as a company, in order to effectively use accumulated information, it is indispensable not only to accumulate valid information but also to systematically organize the accumulated information. Conventionally, a method for sharing information by manually storing information to be shared in a directory corresponding to a category considered to be appropriate for an information sharing system such as a file server has been generally used.
On the other hand, in Patent Document 1, a series of related mails such as reply mails (Re: mails) are extracted from stored e-mails (hereinafter simply referred to as “mails”). Is a system that displays a transmission route for e-mails and information on participants such as senders and recipients. Patent Document 2 describes a system that generates attributes of a person who is a transmission source and a destination from stored mail information.

JP 2002-14903 A JP 2001-175680 A

As described above, it is difficult to classify and store all information in appropriate categories by manually storing information in a directory corresponding to a category considered appropriate in the information sharing system. There is also a possibility that classification cannot be performed well when belonging to a plurality of categories. Even if information is stored in an appropriate category, information such as related persons and related organizations that change with the passage of time is not reflected. Furthermore, in most cases, the accumulated information is not all information related to what is desired to be known. These are the reasons why a person who wants to know a certain search target cannot quickly find information related to the search target. So, if you want to know about a certain search target, finding the person most relevant to the search target and making an inquiry can be said to be a quick and reliable way to obtain information.
However, in the techniques of Patent Documents 1 and 2, information about who is the most relevant person for the search target is not presented to the searcher.

  The present invention has been made to solve the above problems, and a purpose of the present invention is to provide a search support apparatus capable of detecting and presenting a highly relevant person for a search target by using an electronic mail as community information. And a method, and a computer program.

In order to solve the above-described problem, the present invention provides a record including at least message identification information for identifying each electronic mail, thread identification information for identifying a thread to which the electronic mail belongs, and information on a sender and a receiver. And an analysis result storing means for storing a keyword table indicating information in which thread identification information of each thread is associated with a keyword, and a mail storage means for storing the e-mail. Whether or not the message identification information of other related emails is acquired from the information of the email reading unit to be read and the header information of the email read by the email reading unit, and whether the acquired message identification information is registered in the email table E-mail analyzing means for judging If it is determined, the thread identification information of the read e-mail is newly generated, and if it is determined that it is registered, the mail table specified by the message identification information of the other related e-mail Thread identification information is acquired from the record of the message, the message identification information acquired from the header information of the read email, the sender and the receiver, and the newly generated thread identification information or the thread identification information acquired from the mail table A keyword that is a result of morphological analysis of information obtained from the read e-mail, and an analysis result output unit that writes to the keyword table in association with the thread identification information; Receiving means for inputting search conditions, and the keyword table The thread identification information of the keyword that matches the search word indicated by the search condition received by the receiving means is acquired, and for each acquired thread identification information, the sender or the record from the mail table record specified by the thread identification information Search means for acquiring a user set as a receiver, the number of the user set as a sender or receiver, and thread identification information indicating a user whose number acquired by the search means exceeds a predetermined condition Output control means for each output , the mail storage means, the user who is the speaker in the conference as the sender, the participants of the conference as the receiver, the conference name of the conference as the subject, The utterance content of the user who is the speaker is set in the text, and includes information indicating that it was generated based on the conference, and other related As the message identification information of the child mail, an email in which the message identification information of another email in which other utterance contents in the conference are set is stored, and the analysis result output means is based on the conference in the read email If the information indicating that the information is generated is set, the information to be generated based on the conference is further set in the record to be added to the mail table, and the search unit acquires each acquired thread identification For each information, each record set in the mail table specified by the thread identification information is specified, and information indicating that the record is generated based on the conference is set in each specified record. A user who is given a given point to a user who is set as a sender in a record and who is set as a receiver When information lower than the predetermined point is given and information indicating that it was generated based on the conference is not set, it is set as the user and receiver set as the sender in each record A point that is weighted more than the predetermined point is given to the user, and the output control unit outputs, for each thread identification information, information indicating a user whose total of points given by the search unit is greater than a predetermined condition. This is a search support device.

  Further, the present invention is the above-described search support apparatus, wherein the output control unit selects a user whose number acquired by the search unit is greater than a predetermined condition in an order corresponding to the number of the set user. And output for each thread identification information.

  Further, the present invention is the search support apparatus described above, wherein the search means acquires, from the keyword table, thread identification information of a keyword that matches a search word indicated by a search condition received by the reception means, For each acquired thread identification information, a set of users set as a sender and a receiver is acquired from a record in the mail table specified by the thread identification information, and the output control means is acquired by the search means The set of users is output for each thread identification information.

  Further, the present invention is the search support apparatus described above, wherein the search means acquires, from the keyword table, thread identification information of a keyword that matches a search word indicated by a search condition received by the reception means, For each acquired thread identification information, a set of users set as a sender and a receiver is acquired from a record in the mail table specified by the thread identification information, and the output control means is acquired by the search means Based on the set of users, information indicating the correlation between a plurality of users included in the acquired set of users is output for each thread identification information.

  Further, the present invention is the search support apparatus described above, wherein the search means further acquires the number of sets of the user set from the mail table record specified by the thread identification information. The output control means further outputs information corresponding to a set number of users.

  Further, the present invention is the above-described search support device, wherein the mail table record further includes broadcaster information, and the analysis result output means is acquired from the header information of the read electronic mail. A record in which the information of the broadcaster is further set is added to the mail table, and the search means sends a sender to each record of the mail table specified by the thread identification information for each acquired thread identification information. Giving a predetermined point to a user set as a user and a user set as a recipient, and giving a point weighted to a predetermined weight from the predetermined point to a user set as a broadcaster, The output control unit is configured to display information indicating a user whose total points given by the search unit is greater than a predetermined condition as thread identification information. And outputs each, characterized in that.

  Further, the present invention is the search support apparatus described above, wherein the search means is set as a receiver for each record of the mail table specified by the thread identification information for each acquired thread identification information. A predetermined point is given to a user who is set as a sender, a point weighted with a predetermined weight from the predetermined point is given to a user set as a sender, and the output control means is a total of the points given by the search means Is output for each thread identification information, indicating information indicating that there are more users than the predetermined condition.

  Further, the present invention is the search support apparatus described above, wherein the search means is a receiver set in each record of the mail table specified by the thread identification information for each acquired thread identification information. If the address is a mailing list address, a predetermined point is given to the user set as the sender, and if it is not the mailing list address, the user set as the sender and the user set as the receiver are given. Predetermined points are assigned, and the output control means outputs, for each thread identification information, a user whose total number of points given by the search means exceeds a predetermined condition.

Further, the present invention is a mail table comprising message identification information for identifying each electronic mail, thread identification information for identifying a thread to which the electronic mail belongs, and a record including at least information on a sender and a receiver, and Analysis result storage means for storing a keyword table indicating information in which thread identification information of each thread is associated with a keyword, mail reading means for reading an email from a mail storage means for storing the email, and Mail analysis means for acquiring message identification information of other related electronic mails from the information of the header of the electronic mail read by the mail reading means, and determining whether the acquired message identification information is registered in the mail table; If the e-mail analysis means determines that it is not registered, If it is determined that a new email thread identification information is registered and registered, the thread identification information is obtained from the record in the mail table specified by the message identification information of the other related email. And a record in which the message identification information acquired from the header information of the read electronic mail, the sender and the receiver, and the newly generated thread identification information or the thread identification information acquired from the mail table are set in the mail table. And a keyword that is a result of morphological analysis of information obtained from the read e-mail, an analysis result output means that writes the keyword identification information in association with the thread identification information, and a reception that receives input of search conditions Means and the receiving means from the keyword table. The thread identification information of the keyword that matches the search word indicated by the digit search condition is acquired, and for each acquired thread identification information, set as a sender or a receiver from the record of the mail table specified by the thread identification information For each thread identification information, a search means for acquiring the number of users who have been set as a sender or a receiver, and a user whose number acquired by the search means exceeds a predetermined condition Output control means, wherein the record of the mail table further includes information on the broadcaster, and the analysis result output means includes information on the broadcaster obtained from the header information of the read electronic mail. Further, the set record is added to the mail table, and the search means for each acquired thread identification information. A predetermined administrator search corresponding point is given to a user set as a sender and a user set as a receiver in each record of the mail table specified by different information, and set as a broadcaster. A higher administrator search support point than the predetermined administrator search support point, and the output control means displays information indicating a user whose total of the administrator search support points exceeds a predetermined condition as an administrator. The search support apparatus is characterized in that it outputs each thread identification information.
Further, the present invention is the search support apparatus described above, wherein the analysis result storage unit further stores thread identification information and a thread table indicating information on a start date and time of a thread specified by the thread identification information. The analysis result output means associates the thread identification information with the start date and time indicating the date and time information acquired from the header information of the email when the thread identification information of the read email is newly generated. In addition, the search means further reads out information on the start date and time corresponding to the thread identification information acquired from the keyword table, and the output control means reads the start date and time that matches the received period condition The user or the set of users is output for the thread identification information.

  Further, the present invention is the above-described search support device, wherein the record of the mail table further includes date information, and the analysis result output means stores the date acquired from the header information of the read electronic mail. A record in which information is further set is added to the mail table, the search means further acquires date information from the record in the mail table specified by the acquired thread identification information, and the output control means is the search Based on the date information corresponding to each thread identification information acquired by the means, information indicating the number of e-mails for each predetermined period is output for each thread identification information.

  Further, the present invention is the search support apparatus described above, wherein the analysis result output means uses the subject obtained from the read e-mail or the result of morphological analysis of the subject and text information as a keyword. Write to table.

  Further, the present invention is the search support apparatus described above, wherein the analysis result output means writes a result of a morphological analysis with a predetermined part of speech as a keyword in the keyword table.

  Further, the present invention is the search support apparatus described above, wherein the analysis result output means writes a result obtained by removing a predetermined unnecessary character string from the result of morphological analysis as a keyword in the keyword table. It is characterized by that.

  In addition, the present invention provides the above-described search support device, comprising: a person information storage unit that stores a person table indicating information that associates information for identifying a user and the name of the user; and the mail reading unit. From the information in the body text of the read e-mail, extract the line that contains the predetermined word used for self-introduction, and select the proper noun word from the words obtained by morphological analysis of the extracted line And a person information generating means for associating the selected word and the information for identifying the sender acquired from the read header information of the electronic mail and writing it in the person table, wherein the output control means outputs The name of the user read from the person table corresponding to the information specifying the target user or the information specifying the users constituting the set of users to be output Output, characterized in that.

  Also, the present invention is the search support apparatus described above, wherein the person information generating means is a predetermined number used for self-introduction from a line up to a predetermined number of lines from the head set in the information of the text of the electronic mail. It is characterized in that a line including the word is extracted.

  The present invention is also the search support apparatus described above, wherein the person information generating means extracts a phrase before or after a delimiter from a line extracted as including a predetermined phrase used for self-introduction. Then, morphological analysis is performed, and the phrase obtained as the proper noun is written in the person table in association with the information specifying the sender acquired from the read header information of the electronic mail.

  Further, the present invention is the above-described search support device, wherein the person information generating means uses the sender's email address acquired from the header information of the read email as information for identifying the sender. It is characterized by writing in the person table.

  Further, the present invention is the above-described search support device, wherein the search means includes information indicating that the data generated based on the audio of the conference is set in each of the specified records. The predetermined point is given only to a user set as a sender in each record.

  Further, the present invention is the search support apparatus described above, wherein the search means acquires, from the keyword table, thread identification information of a keyword that matches a search word indicated by a search condition received by the reception means, For each acquired thread identification information, a set of users set as a sender and a receiver and the number of sets of the set of users are acquired from the mail table record specified by the thread identification information. When the information indicating that the record is generated based on the conference is set in the record, a predetermined point is given to the set of each user according to the number of the set of the user, When information indicating that the record is generated based on the conference is not set, the set of each user is weighted from the predetermined point. The poppy points assigned depending on the number which is set set of the user, the output control means outputs information corresponding to the set point of each user, and wherein the.

  Further, the present invention is the search support apparatus described above, wherein the analysis result output means starts from the read e-mail when information indicating that the read e-mail is generated based on the conference is set. The result of morphological analysis of the obtained information is used as a keyword candidate, and when the keyword candidate is obtained a predetermined number of times or more from an email corresponding to the same thread identification information, the keyword candidate is associated with the thread identification information. It writes to the said keyword table, It is characterized by the above-mentioned.

  The present invention is also the search support apparatus described above, wherein the audio data storage means stores conference data including a conference name, audio data of audio during the conference, and participants indicating users who have participated in the conference; Utterer data storage means for storing voice feature information, which is a parameter indicating the voice characteristics of each user, a conference file is read from the voice data storage means, and the voice data in the read conference file is stored as one user. The voice data dividing means for generating the divided voice data divided into the speech sections in which the voice is spoken, the divided voice data, and the voice feature information of each user stored in the speaker data storage means, Utterer recognition means for identifying a certain user, speech text conversion means for generating text data in which the speech content is converted into text from the divided speech data, The text specified by the spoken word recognition means is sent to the sender, the participant in the conference data is the recipient, the name of the meeting in the conference data is the subject, and the text data generated by the speech text conversion means is the body text. To the other divided voice data extracted from the same voice data as newly generated message identification information, information indicating that it was generated based on the meeting, and message identification information of other related emails And a mail conversion unit that generates an e-mail in which the message identification information of the e-mail generated based on the e-mail is set and writes the generated e-mail in the mail storage unit.

  Further, the present invention is the search support apparatus described above, wherein the conference data further includes information on a meeting date and time, and the audio data dividing unit corresponds to the audio indicated by the divided audio data. An elapsed time from the beginning of the voice indicated by the data is acquired, and the mail converting means is an electronic device further setting a date and time obtained by adding the elapsed time acquired by the voice data dividing means to the meeting date and time in the meeting data. Generate mail

Further, the present invention is a mail table comprising message identification information for identifying each electronic mail, thread identification information for identifying a thread to which the electronic mail belongs, and a record including at least information on a sender and a receiver, and A search support apparatus comprising analysis result storage means for storing a keyword table indicating information in which thread identification information of each thread is associated with a keyword, and presenting an important person related to desired information to assist the search of the information The mail reading means reads the e-mail from the mail storage means storing the e-mail, and the mail analyzing means uses the e-mail header information read by the mail reading means. To obtain the message identification information of other related e-mails from Is registered in the mail table, and if the analysis result output means determines that it is not registered by the mail analysis means, it newly generates thread identification information of the read electronic mail. If it is determined that the message is registered, the thread identification information is obtained from the record of the mail table specified by the message identification information of the other related email, and is obtained from the header information of the read email. A record in which message identification information, sender and receiver, and newly generated thread identification information or thread identification information acquired from the mail table is set is added to the mail table and obtained from the read e-mail. The keyword, which is the result of morphological analysis of the information, is associated with the thread identification information before Writing to the keyword table, the accepting means receives the input of the search condition, the search means obtains the thread identification information of the keyword that matches the search word indicated by the search condition accepted by the accepting means from the keyword table, For each acquired thread identification information, the number of users set as senders or receivers from the mail table record specified by the thread identification information, and the number of users set as senders or receivers And the output control means outputs a user whose number obtained by the search means is greater than a predetermined condition for each thread identification information , and the mail storage means is a user who is a speaker in a meeting. The sender, the participant of the conference as the receiver, the conference name of the conference as the subject, and the user who is the speaker. The other utterance contents in the conference are set as the message identification information of other related e-mails including information indicating that the utterance contents of the user are set in the text and generated based on the meeting. The e-mail storing the message identification information of the e-mail is stored, and the analysis result output means adds to the e-mail table when information indicating that the read e-mail is generated based on the meeting is set. In the record, the information indicating that it is generated based on the conference is further set, and the search means sets each record set in the mail table specified by the thread identification information for each acquired thread identification information. If the information indicating that the record was generated based on the meeting is set in each specified record, the sender Is given to the user who is set as the recipient, and the user who is set as the recipient is given a point lower than the predetermined point, and information indicating that the user has been generated based on the conference is set. If not, the user who is set as the sender and the user who is set as the receiver in each record is given a point weighted more than the predetermined point, and the output control means is given by the search means The search support method is characterized in that information indicating a user whose total number of points is greater than a predetermined condition is output for each thread identification information .
Further, the present invention is a mail table comprising message identification information for identifying each electronic mail, thread identification information for identifying a thread to which the electronic mail belongs, and a record including at least information on a sender and a receiver, and A search support apparatus comprising analysis result storage means for storing a keyword table indicating information in which thread identification information of each thread is associated with a keyword, and presenting an important person related to desired information to assist the search of the information The mail reading means reads the e-mail from the mail storage means storing the e-mail, and the mail analyzing means uses the e-mail header information read by the mail reading means. To obtain the message identification information of other related e-mails from Is registered in the mail table, and if the analysis result output means determines that it is not registered by the mail analysis means, it newly generates thread identification information of the read electronic mail. If it is determined that the message is registered, the thread identification information is obtained from the record of the mail table specified by the message identification information of the other related email, and is obtained from the header information of the read email. A record in which message identification information, sender and receiver, and newly generated thread identification information or thread identification information acquired from the mail table is set is added to the mail table and obtained from the read e-mail. The keyword, which is the result of morphological analysis of the information, is associated with the thread identification information before Writing to the keyword table, the accepting means receives the input of the search condition, the search means obtains the thread identification information of the keyword that matches the search word indicated by the search condition accepted by the accepting means from the keyword table, For each acquired thread identification information, the number of users set as senders or receivers from the mail table record specified by the thread identification information, and the number of users set as senders or receivers And the output control means outputs a user whose number obtained by the search means is greater than a predetermined condition for each thread identification information, and the mail table record further includes a broadcaster's record. The analysis result output means includes the information of the broadcaster acquired from the header information of the read e-mail. A record in which information is further set is added to the mail table, and the search means is set as a sender in each record of the mail table specified by the thread identification information for each acquired thread identification information. A predetermined administrator search support point is given to a user set as a user and a recipient, and an administrator search support point higher than the predetermined administrator search support point is set to a user set as a broadcaster In the search support method, the output control means outputs, for each thread identification information, information indicating, as an administrator, a user whose total number of administrator search support points is greater than a predetermined condition.

In addition, the present invention provides a computer used as a search support apparatus that presents an important person related to desired information and supports the search of the information, message identification information for identifying each electronic mail, and a thread to which the electronic mail belongs Analysis that stores a thread identification information that identifies a message, a mail table that includes at least a record including information on the sender and the receiver, and a keyword table that indicates information that associates the thread identification information of each thread with a keyword. Result storage means, two-stage mail readers for reading e-mails from mail storage means for storing e-mails, and other related e-mail messages from information on headers of e-mails read by the e-mail reading means Acquires identification information, and the acquired message identification information is registered in the mail table. If it is determined by the mail analysis means that determines whether or not it is registered, it is determined that the thread identification information of the read email is newly generated and registered. In this case, the thread identification information is acquired from the record of the mail table specified by the message identification information of the other related electronic mail, the message identification information acquired from the header information of the read electronic mail, the sender and the receiver And a keyword that is a result of morphological analysis of information obtained from the read e-mail while adding a record in which the newly generated thread identification information or thread identification information acquired from the mail table is set to the mail table And the amount to be written to the keyword table in association with the thread identification information From the result output means, the receiving means that receives the input of the search condition, and the keyword table, the thread identification information of the keyword that matches the search word indicated by the search condition received by the receiving means is acquired, and for each acquired thread identification information Search means for acquiring a user set as a sender or a receiver from the record of the mail table specified by the thread identification information, and a number of the user set as a sender or a receiver, The mail storage means functions as an output control means for outputting, for each thread identification information, a user whose number obtained by the search means is greater than a predetermined condition , and the mail storage means uses the user who is a speaker in the conference as the sender. With the participants as recipients, the meeting name of the conference as the subject, and the utterance content of the user who is the speaker Message identification information of other e-mails set with other utterance contents in the conference as message identification information of other related e-mails The set e-mail is stored, and when the information indicating that the analysis result output means is generated based on the meeting is set in the read e-mail, the record added to the e-mail table includes the meeting The information indicating that the information is generated based on the thread identification information is further set, for each thread identification information obtained, the search means identifies each record set in the mail table identified by the thread identification information, If information indicating that the record was generated based on the conference is set for each identified record, the sender is set for each record. If a user is given a predetermined point and a point lower than the predetermined point is given to a user who is set as a recipient, and information indicating that it has been generated based on a conference is not set, A point weighted more than the predetermined point is given to the user set as the sender and the user set as the receiver in each record, and the output control means is the sum of the points given by the search means Is a computer program that outputs information indicating users who have more than a predetermined condition for each thread identification information .
In addition, the present invention provides a computer used as a search support apparatus that presents an important person related to desired information and supports the search of the information, message identification information for identifying each electronic mail, and a thread to which the electronic mail belongs Analysis that stores a thread identification information that identifies a message, a mail table that includes at least a record including information on the sender and the receiver, and a keyword table that indicates information that associates the thread identification information of each thread with a keyword. Result storage means, mail reading means for reading an e-mail from a mail storage means for storing the e-mail, and message identification information of other related e-mails from information on the header of the e-mail read by the mail reading means And the acquired message identification information is registered in the mail table. Mail analysis means for determining whether or not it is registered by the mail analysis means, when it is determined that the thread identification information of the read email is newly generated and registered Obtains the thread identification information from the mail table record specified by the message identification information of the other related electronic mail, and the message identification information obtained from the header information of the read electronic mail, the sender and the receiver Adding a record in which the newly generated thread identification information or the thread identification information acquired from the mail table is set to the mail table, and a keyword that is a result of morphological analysis of information obtained from the read email; , Analysis that writes to the keyword table in association with the thread identification information From the result output means, the receiving means for receiving the input of the search condition, the thread identification information of the keyword that matches the search word indicated by the search condition received by the receiving means is acquired from the keyword table, and for each acquired thread identification information Search means for acquiring a user set as a sender or a receiver from the record of the mail table specified by the thread identification information, and a number of the user set as a sender or a receiver, The number of users acquired by the search means is functioned as output control means for outputting a user whose number is greater than a predetermined condition for each thread identification information, and the mail table record further includes information on the broadcaster, The output means further sets the broadcaster information acquired from the header information of the read email A record is added to the mail table, and for each acquired thread identification information, the search means, as a user and a receiver set as a sender in each record of the mail table specified by the thread identification information A predetermined administrator search corresponding point is given to the set user, and an administrator search corresponding point higher than the predetermined administrator search corresponding point is given to the user set as a broadcaster, and the output control The means is a computer program that outputs, for each thread identification information, information indicating, as an administrator, a user whose total number of points corresponding to the administrator search exceeds a predetermined condition.

  According to the present invention, it is possible to present an important person for a certain search target, and to grasp the connection between important persons. In addition, for example, in the case of system development, related persons and organizations change as development progresses, such as a specification review stage, a program design stage, a test stage, and an introduction stage. It is possible to grasp the flow of presenting important persons and changing important persons and organizations.

It is a functional block diagram of the search assistance apparatus by the 1st Embodiment of this invention. It is a data block diagram of the mail table by the embodiment. It is a data block diagram of the thread table by the embodiment. It is a data block diagram of the keyword table by the embodiment. It is a figure which shows the analysis processing flow of the mail collection part by the embodiment. It is an example of a mail header. It is a figure which shows the example of a setting of the mail table by the embodiment. It is a figure which shows the example of a setting of the thread table by the embodiment. It is a figure which shows the example of a setting of the keyword table by the embodiment. It is a figure which shows the thread | sled search process flow of the search process part by the embodiment. It is a figure which shows the person correlation table creation process flow of the search process part by the embodiment. It is a figure which shows the important person table creation process flow of the search process part by the embodiment. It is a figure which shows the result display processing flow of the output control part by the embodiment. It is a figure which shows the example of a setting of the mail table by the embodiment. It is a figure which shows the example of a setting of the thread table by the embodiment. It is a figure which shows the example of a setting of the keyword table by the embodiment. It is a figure which shows the example of a setting of the person correlation table by the same embodiment, and an important person table. It is a figure which shows the output image of the important person by the embodiment. It is a figure which shows the display image of the important person at the time of search time designation | designated by the same embodiment, and a person correlation diagram. It is a figure which shows the whole image of the search screen by the embodiment. It is a figure which shows the thread | sled output image by 2nd Embodiment. It is a figure which shows the thread | sled output image by the embodiment. It is a figure which shows the data block diagram of the mail table by 3rd Embodiment. It is a figure which shows the display image of the person correlation diagram by the embodiment. It is a figure which shows the data block diagram of the mail table by 4th Embodiment. It is a functional block diagram of the search assistance apparatus by the 6th Embodiment of this invention. It is a data block diagram of the person table by the embodiment. It is a figure which shows the person information generation process flow of the person information generation part by the embodiment. It is an example of mail. It is a figure which shows the example of a setting of the mail table by the embodiment. It is a figure which shows the example of a setting of the thread table by the embodiment. It is a figure which shows the example of a setting of the keyword table by the embodiment. It is a figure which shows the example of a setting of the person table by the embodiment. It is a figure which shows the result display processing flow of the output control part by the embodiment. It is a figure which shows the example of a setting of the mail table by the embodiment. It is a figure which shows the example of a setting of the thread table by the embodiment. It is a figure which shows the example of a setting of the keyword table by the embodiment. It is a figure which shows the example of a setting of the person table by the embodiment. It is a figure which shows the example of a setting of the person correlation table by the same embodiment, and an important person table. It is a figure which shows the output image of the important person by the embodiment. It is a figure which shows the display image of the important person at the time of search time designation | designated by the same embodiment, and a person correlation diagram. It is a figure which shows the whole image of the search screen in case there is no person name display by the embodiment. It is a figure which shows the whole image of the search screen in the case of person name display with the embodiment. It is a functional block diagram of the search assistance apparatus by 7th Embodiment. 3 is a functional block diagram of a voice recognition unit according to the embodiment. FIG. It is a figure which shows the mail format data generation processing flow of the speech recognition part by the embodiment. It is a figure which shows the analysis processing flow of the mail collection part by the embodiment. It is a figure which shows the person correlation table creation process flow of the search process part by the embodiment. It is a figure which shows the important person table creation process flow of the search process part by the embodiment.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

[First Embodiment]
FIG. 1 is a functional block diagram of a search support apparatus 1 according to the first embodiment of the present invention, in which only functional blocks related to the present invention are extracted and shown.
In the figure, the search support apparatus 1 can be composed of one or a plurality of servers, etc., and includes a mail storage unit 10, a mail collection unit 20, an analysis result storage unit 30, a search processing unit 40, and output control. Part 50.

The mail storage unit 10 can be realized by, for example, a mail server and stores data of electronic mail (hereinafter also simply referred to as “mail”) transmitted and received between users.
The mail collection unit 20 includes a mail reading unit 21 that reads e-mail data from the mail storage unit 10, a mail analysis unit 22 that analyzes the e-mail data read by the mail reading unit 21, and a mail analysis unit 22. An analysis result output unit 23 is provided for writing data indicating the analysis result of the above to the analysis result storage unit 30.
The analysis result storage unit 30 stores a mail table indicating information about each mail, a thread table indicating the start and end date and time of each thread, and a keyword table indicating a keyword for each thread as mail analysis result data. A thread indicates a group of a series of emails that have been replied or forwarded.

The search processing unit 40 can be realized by a database, for example, and inputs search conditions from an input means provided in the search support device 1 or a computer terminal (output control unit 50) connected to the search support device 1. The received query receiving unit 41, the search unit 42 that searches the mail analysis result data stored in the analysis result storage unit 30 based on the search condition received by the query receiving unit 41, and the search result by the search unit 42 And a search result output unit 43 for outputting.
The output control unit 50 can be realized by, for example, a computer terminal. The search result reading unit 51 that reads the search result output by the search result output unit 43 of the search processing unit 40 and the search result reading unit 51 read the search result. And a result display unit 52 for displaying the search results on the display. The search result may be printed by a printer or written to a recording medium.

Next, each data table stored in the analysis result storage unit 30 will be described.
FIG. 2 is a diagram showing the configuration of the mail table stored in the analysis result storage unit 30. As shown in FIG. In the figure, the mail table is composed of a plurality of records including mail ID, thread ID, message ID, subject, sender, receiver, date, and body data. The mail ID is identification information for uniquely identifying each record in the mail table. The mail analysis unit for each record such that the first line is mail ID = 1, the second line is mail ID = 2,. 22. The thread ID is identification information for uniquely identifying each thread, and is assigned to each thread by the mail analysis unit 22 such that the first thread is thread ID = 1, the second thread is thread ID = 2, and so on. Is done. The message ID is identification information that uniquely identifies each mail, and corresponds to information set in the message-ID field of the mail. The subject, sender, recipient, date, and text correspond to the information set in the email Subject field, From field, To field, Date field, and mail text, respectively.

  FIG. 3 is a diagram illustrating a configuration of a thread table stored in the analysis result storage unit 30. In the figure, the thread table is composed of a plurality of records including thread ID, start date / time, and end date / time data. Here, the start date / time and the end date / time are set to the date obtained from the mail in which the information of the oldest date is set and the information of the latest date among the mails belonging to the thread specified by the thread ID. Indicates the date obtained from the mail.

  FIG. 4 is a diagram showing the configuration of the keyword table stored in the analysis result storage unit 30. As shown in FIG. In the figure, the keyword table is composed of a plurality of records including a record ID, a thread ID, and keyword data. The record ID is identification information that uniquely identifies each record in the keyword table, and is assigned to each record by the mail analysis unit 22. The keyword ID is information obtained as a result of morphological analysis of the subject of the mail belonging to the thread specified by the thread ID.

Next, a processing flow in the search support apparatus 1 will be described.
FIG. 5 shows an analysis processing flow in the mail collection unit 20.
In the figure, the mail reading unit 21 of the mail collection unit 20 reads a mail that has not been analyzed by the present analysis processing flow from the mail server as the mail storage unit 10 (step S110). The mail includes fields such as Message-ID, Subject, From, To, CC, Date, References, and mail text. Note that References is a field that is included only in the mail that has been replied or transferred, and is a field to which the Message-ID that is the reply or transfer source is added each time a reply or transfer is made.

  When the mail analysis unit 22 assigns a new natural mail ID, the analysis result output unit 23 adds the record in which the mail ID is set to the mail table stored in the analysis result storage unit 30. The analysis result output unit 23 adds the message-ID, Subject, From, To, and Date of the read mail to the message ID, subject, sender, recipient, date, and body data items of the added record. Then, the setting contents of the mail body field are written (step S120).

  Next, the mail analysis unit 22 searches the mail table, and determines whether there is a record in which data having the same contents as the data set in the Reference field of the read mail is set in the message ID (Step S1). S130). Thereby, it is confirmed whether or not the read mail belongs to the same thread as the previous mail. When it is determined that there is no record in which the data having the same content as the Reference field of the read mail is set in the message ID (step S130: NO), the mail analysis unit 22 determines that the new thread is a new thread ID. Is granted. The analysis result output unit 23 adds the record in which the newly assigned thread ID is set to the thread table stored in the analysis result storage unit 30 (step S140), and the start date / time and end date / time of the added record. Data set in the date field of the read mail is written in the data item (step S150).

  On the other hand, if it is determined that there is a record in which the same data as the Reference field of the read mail is set in the message ID (step S130: YES), the possibility of the last mail belonging to the already registered thread Judge that there is. Therefore, the mail analysis unit 22 reads the thread ID from the record in which the data having the same contents as the reference field of the read mail is set in the message ID, and searches the thread table by this thread ID. Then, the date and time set in the start date and time and the end date and time of the detected record are compared with the date and time set in the Date field of the read mail. If the date and time set in the Date field of the read email is older than the start date and time, the data item of the start date and time of the detected record is set to the date and time set in the Date field of the read email rewrite. Also, if the date and time set in the Date field of the read email is newer than the end date and time, the data item of the end date and time of the detected record is set in the Date field of the read email The date is rewritten (step S160).

  The analysis result output unit 23 newly writes the thread ID of the mail in the mail table record added in step S120 (step S170). Next, the mail analysis unit 22 performs morphological analysis on the setting contents of the subject field of the mail and assigns a new record ID to each keyword that is a result of the morphological analysis. The analysis result output unit 23 adds a record including a newly assigned record ID, a thread ID of the mail, and a keyword corresponding to the record ID to the keyword table stored in the analysis result storage unit 30 ( Step S180). The mail reading unit 21 determines whether there is a mail that has not yet been analyzed in the mail server as the mail storage unit 10 (step S190). If there is an email that has not been analyzed yet (step S190: YES), the process returns to step S110, and the above-described processing is repeated. Then, when there is no mail that has not been analyzed in the mail storage unit 10 (step S190: NO), the process ends.

The process of FIG. 5 will be described using a specific data example.
6 shows an example of the mail header of each mail read from the mail storage unit 10, FIG. 7 shows an example of a mail table generated from the mail having the mail header shown in FIG. 6, and FIG. 9 shows an example of a thread table generated from a mail having a mail header. FIG. 9 shows an example of a keyword table generated from a mail having a mail header shown in FIG.

  First, when the mail M1 (FIG. 6) is read (FIG. 5, step S110), a record in which the newly assigned mail ID “1” is set is added to the mail table (FIG. 7). message-ID "1111@svr.jp", Subject "Regarding development progress", From "userA", To "userB", Date "2007/2/3 15:12", the email body is the message ID of the record, subject , Sender, receiver, date, and text are set (FIG. 5, step S120). Further, since the reference field is not included in the mail M1, it is determined that the mail is a new thread (FIG. 5, step S130: NO). Therefore, a record in which a new thread ID “1” is set is added to the thread table (FIG. 8) (FIG. 5, step S140), and the date “2007/2/3 15:12” of the mail M1 is the start of the record. The date / time and end date / time are set (FIG. 5, step S150), and the thread ID “1” is set in the record of the mail ID “1” in the mail table (FIG. 7) (FIG. 5, step S170). Further, when “development”, “progress”, and “about” are obtained by performing morphological analysis on the setting contents of the subject field of the mail M1, newly assigned record IDs “1”, “ Records set with “2” and “3” are added to the keyword table (FIG. 9). Each of the added records includes a thread ID “1” and a corresponding morphological analysis result “development”, “progress”, “About” is set (FIG. 5, step S180).

  Next, when the mail M2 (FIG. 6) is read (FIG. 5, step S110), a record in which the newly assigned mail ID “2” is set is added to the mail table (FIG. 7). message-ID "2222@svr.jp", Subject "Re: Regarding development progress", From "userB", To "userA", Date "2007/2/4 0:12", the mail body is the message ID of the record , Subject, sender, receiver, date, and text are set (FIG. 5, step S120). Since there is already a record in the mail table in which the reference “1111@svr.jp” of the mail M2 is set as the message ID (record with the mail ID “1”), it is determined that the thread has already appeared ( FIG. 5, Step S130: NO). Since the thread ID “1” is set in this record, the start date and time “2007/2/3 15:12” and the end date and time of the record in which the thread ID is set are shown in the thread table (FIG. 8). “2007/2/3 15:12” is read and compared with the date “2007/2/4 0:12” of the mail M2. Since the date of the mail M2 is newer than the end date and time, the data item of the end date and time is rewritten to the setting contents of the date of the mail M2 (step S160), and the record of the mail ID “2” in the mail table (FIG. 7) Is set with the thread ID “1” (FIG. 5, step S170). Furthermore, “development”, “progress”, and “related” are obtained by performing morphological analysis on the setting contents of the Subject field of the mail M1, but as shown in FIG. 9, the keyword table (FIG. 9) includes the relevant thread. Since these keywords are already set corresponding to the ID “1”, no new record is added.

  The result of performing the processing of FIG. 5 on the mails M1 to M8 shown in FIG. 6 is the mail table shown in FIG. 7, the thread table shown in FIG. 8, and the keyword table shown in FIG. As shown in FIG. 7, records are added to the mail table for each of the mails M1 to M8, the thread ID “1” is assigned to the records corresponding to the mails M1, M2, and M3, and the record corresponding to the mail M4 is assigned to the mail table. A thread ID “3” is set for a record corresponding to the mail M5, M6, and M7, and a thread ID “4” is set for a record corresponding to the mail M8. In this case, References “1111@svr.jp” in the mail M2 is set to the message ID of the record corresponding to the mail M1, and References “1111@svr.jp” and “2222@svr.jp” in the mail M3. However, since the message IDs of the records corresponding to the mails M1 and M2 are set, the same thread ID is set to the records corresponding to the mails M1, M2, and M3. Similarly, References “5555@svr.jp” of the mail M6 is set to the message ID of the record corresponding to the mail M5, and References “5555@svr.jp” and “6666@svr.jp” of the mail M7. Is set to the message ID of the record corresponding to the mail M5 and mail M6, the same thread ID is set to the record corresponding to the mail M5, M6 and M7.

Further, in the thread table shown in FIG. 8, in the record of the thread ID “1”, the setting content of the Date field of the first mail M1 belonging to the thread ID is the start date and time of the last mail M3 belonging to the thread ID. The setting content of the Date field is set to the end date and time. Similarly, in the record of the thread ID “3”, the setting contents of the Date field of the first mail M5 belonging to the thread ID are the start date and time, and the setting contents of the Date field of the last mail M7 belonging to the thread ID are the end date and time. Is set to
Since only the mail M4 belongs to the thread ID “2”, the setting contents of the Date field of the mail M4 are set to the start date and time and the end date and time in the record of the thread ID. Similarly, in the record with the thread ID “4”, the setting contents of the Date field of the mail M8 are set to the start date and time and the end date and time.

  Further, as shown in FIG. 9, the keyword table includes “development”, “progress”, “related” as a result of morphological analysis of the subject setting contents of the mails M1, M2, and M3 belonging to the thread ID “1”. “Social gathering” as a result of morphological analysis of the subject settings of the mail M4 belonging to the thread ID “2” in the record IDs “1” to “3” belongs to the record “4” and the thread ID “3”. Subject settings for email M5, M6, and M7: “Meetings” and “results” as a result of morphological analysis. Subject settings for email M8 belonging to record ID “5” and “6” and thread ID “4” “Development” and “Conference” as a result of the morphological analysis are set to record IDs “7” and “8”.

10 to 12 show a processing flow in the search processing unit 40.
FIG. 10 shows a thread search processing flow in the search processing unit 40.
The query receiving unit 41 of the search processing unit 40 reads a search word from the received search request (step S210). The search unit 42 selects and reads a thread ID that has not been subjected to the thread search process according to this thread search process flow from among the thread IDs set in the keyword table stored in the analysis result storage unit 30 (Step S42). S220). Then, it is determined whether or not there is a record in which a keyword matching the search word is set among the records in the keyword table in which the selected thread ID is set (step S230).

  If there is a record in which a keyword that matches the search word is set in the record in which the selected thread ID is set (step S230: YES), the following steps S240 to S250 are executed. That is, the search unit 42 searches the thread table stored in the analysis result storage unit 30 using the selected thread ID as a key, reads the start date and time from the detected record (step S240), and the selected thread ID and Information associated with the read start date and time is stored (step S250).

  If there is a thread ID that has not yet been subjected to thread search processing in the keyword table (step S260: NO), the search unit 42 repeats the processing from step S220. If thread search processing has been executed for all thread IDs registered in the keyword table (step S260: YES), the processing ends.

FIG. 11 shows a person correlation table creation process flow in the search processing unit 40.
The search unit 42 of the search processing unit 40 has not yet performed the person correlation table creation process by the person correlation table creation process flow among the thread IDs stored in step S250 of FIG. 10 (hereinafter, the process target thread ID). Select an ID. The search unit 42 searches the mail table, identifies and reads the record in which the selected thread ID is set (step S310).

  The search unit 42 extracts a set of users set as a sender and a receiver as a set of senders and receivers from the record read in step S310, and calculates the number extracted for each of the extracted sets of senders and receivers. Count (step S320). In addition, even if the sender and the user of the receiver are switched, it is assumed that they are the same set. When the processing target thread ID includes a thread ID that has not yet been subjected to the person correlation table creation process (step S330: NO), the search unit 42 repeats the process from step S310. When the person correlation table creation process is executed for all process target thread IDs (step S330: YES), for each thread ID, the pair of senders and receivers extracted in step S320 and the number of extractions of the pair are supported. A person correlation table, which is a list showing the points, is generated and stored (step S340). That is, each time a user becomes a mail sender or receiver, 1 point is given to the user.

FIG. 12 shows an important person table creation processing flow in the search processing unit 40.
The search unit 42 of the search processing unit 40 selects a thread ID that has not yet been subjected to the important person table creation process in the important person table creation process flow from the process target thread IDs stored in step S250 of FIG. . The search unit 42 searches the mail table, identifies and reads the record in which the selected thread ID is set (step S410).

  The search unit 42 extracts a user set as a sender or a receiver from the read record, and measures the number of appearances that is the number of the extracted user set as a sender or a receiver. (Step S420). If there is a thread ID that has not yet undergone the important person table creation process in the process target thread ID (step S430: NO), the search unit 42 repeats the process from step S410. If the important person table creation process is executed for all process target thread IDs (step S430: YES), for each thread ID, the user extracted in step S420 and the points corresponding to the number of appearances are displayed. An important person table which is a list to be shown is generated and stored (step S440). That is, each time a user becomes a mail sender or receiver, 1 point is given to the user.

FIG. 13 shows a result display processing flow of the output control unit 50.
In the figure, the search result reading unit 51 includes the start date / time of each processing target thread ID output from the search result output unit 43, the person correlation table, the important person table, and the mail table corresponding to each processing target thread ID. A list of subject names read from the record is read as a search result (step S510). Next, the result display unit 52 reads information on the search result display method (step S520). This is information set in the search request received by the query reception unit 41 or stored in advance in the storage unit of the output control unit 50. If it is set in the search result display method information that important person display is required (step S530: YES), the result display unit 52 points out of the users shown in the important person table for each thread. A predetermined number of users are extracted as important persons in descending order. Then, the extracted important person of each thread is displayed together with the number of points corresponding to the important person (step S540). At this time, threads are arranged in time series, and important persons are displayed in the order of the most appearances corresponding to each thread. In addition, when there is an input to display only a specific thread, only the important person for the specified thread is displayed. Display important people for threads whose start date is included in the period.

  Subsequently, when it is set in the search result display method information that the theme name needs to be displayed (step S550: YES), the result display unit 52 reads from the search result output unit 43 in step S510. A list of subject names is displayed (step S560). At this time, the subject may be displayed in association with the display of the important person of each thread output in step S540. When information indicating a period is input, a subject for a thread whose start date and time is included in the period is displayed.

  Furthermore, when it is set in the search result display method information that it is necessary to display the correlation diagram of the heavy person (step S570: YES), the result display unit 52 uses the person correlation table for each thread. A predetermined number of sets are extracted in descending order of the number of points from the set of senders and receivers shown. Then, the sender / receiver pair extracted for each thread is displayed together with the number of points corresponding to the number of points of the sender / receiver pair (step S580). In addition, at this time, it displays in the matrix form which shows the correlation of the group of transmission / reception between several users. In addition, when there is an instruction to display only a specific thread, only the set of senders and receivers for the specified thread is displayed, and information indicating the period is input. , A set of senders and receivers for a thread whose start date and time is included in the period is displayed.

The processes of FIGS. 10 to 13 will be described using specific data examples.
FIG. 14 shows an example of a mail table, FIG. 15 shows an example of a thread table, and FIG. 16 shows an example of a keyword table, which are set before the query receiving unit 41 reads a search keyword. Further, it is assumed that “sales” is input as a search keyword. FIG. 17 shows the person correlation table and the important person table created at this time.

  In the thread search process of FIG. 10, when a search keyword is read (FIG. 10, step S210), thread IDs set in the keyword table (FIG. 16) are selected in order (FIG. 10, step S220). When the thread ID “116” is selected, a record (record ID “340”) in which the search word “sales” is set as a keyword is recorded from a record in the keyword table (FIG. 16) in which the thread ID is set. It is detected (FIG. 10, step S230). The start date and time “2006/7/1 15:12” corresponding to the thread ID is read and stored in association with the thread ID “116” (FIG. 10, steps S240 and S250).

  Next, when the thread ID “120” is selected, the record (record ID “372”) in which the search word “sales” is set as a keyword from the record of the keyword table (FIG. 16) in which the thread ID is set. ") Is detected (FIG. 10, step S230). Then, the start date and time “2007/1/7 13:10” corresponding to the thread ID is read and stored in association with the thread ID “120” (FIG. 10, steps S240 and S250).

Next, when the thread ID “144” is selected, it is detected that there is no record in which the search word “sales” is set as a keyword in the keyword table (FIG. 16) in which the thread ID is set. (FIG. 10, step S230).
As a result, the processing target thread IDs “116” and “120” and the start dates and times “2006/7/1 15:12” and “2007/1/7 13:10” are stored (FIG. 10, step). S250).

Subsequently, a person correlation table creation process (FIG. 11) is started, and process target thread IDs are sequentially selected. When the thread ID “116” is selected, the records (mail IDs “500” to “506”) of the mail table (FIG. 14) in which the thread ID is set are specified (FIG. 11, step S310). As shown in FIG. 17, from the information of the sender and receiver of the identified record, “userA and userB”, “userA and userC”, “userA and userD”, “userA and userE”, “userB and userE” ”Are extracted, and the numbers of these extracted groups are“ 4 ”,“ 2 ”,“ 2 ”,“ 2 ”, and“ 3 ”, respectively (FIG. 11, step S320). In the drawings, userA, userB, userC, userD, and userE may be simply expressed as A, B, C, D, and E.
Next, when the thread ID “120” is selected, the records (mail IDs “507” to “511”) of the mail table (FIG. 14) in which the thread ID is set are specified (FIG. 11, step S310). ). As shown in FIG. 17, a set of “userA and userB”, “userA and userC”, “userA and userD”, and “userA and userE” are extracted from the sender and receiver information of the specified record. The numbers from which these sets are extracted are “2”, “2”, “3”, and “1”, respectively (FIG. 11, step S320).
As a result, for each thread ID, a person correlation table (FIG. 17) is generated that shows these extracted user sets and points corresponding to the number of extracted sets (FIG. 11, step S340). Here, the extraction number 1 is 1 point.

Subsequently, the important person table creation process (FIG. 12) is started, and the process target thread IDs are sequentially selected. When the thread ID “116” is selected, the records (mail IDs “500” to “506”) of the mail table (FIG. 14) in which the thread ID is set are specified (FIG. 12, step S410). As shown in FIG. 17, “userA”, “userB”, “userC”, “userD”, and “userE” are extracted as the users set as the sender or receiver of the specified record, The number of appearances “10”, “7”, “2”, “2”, “5” of the extracted user is measured.
Next, when the thread ID “120” is selected, the records (mail IDs “507” to “511”) of the mail table (FIG. 14) in which the thread ID is set are specified (FIG. 12, step). S420). As shown in FIG. 17, “userA”, “userB”, “userC”, “userD”, and “userE” are extracted as the users set as the sender or receiver of the specified record, The number of appearances “8”, “2”, “2”, “3”, “1” of the extracted user is measured.
Thereby, an important person table indicating these extracted users and points corresponding to the number of appearances is generated for each thread ID (FIG. 12, step S440). Here, the number of appearances is 1 point.

FIG. 18 is an important person output image based on the important person table shown in FIG. 17 and shows an image in which threads are arranged in time series. Here, there are three important persons in descending order of the number of points. Note that the number of points may be equal to or greater than a predetermined number, and the condition is stored in advance in the output control unit 50 or input by the user.
In FIG. 18, the thread IDs “116” and “120” included in the important person table (FIG. 17) are assigned the start dates and times “2006/07/01 15:12” and “2007/07/08 23:11”. On the time axis corresponding to “”. For the thread ID “116”, “userA”, “userB”, “userE” are displayed in descending order of the number of points, and for the thread ID “120”, “userA”, “userD” are displayed in descending order of the number of points. , “UserB”, “userC” are displayed. Since “userB” and “userC” have the same number of points, both are displayed.

FIG. 19 shows a display image of the important person and the person correlation diagram when the search time is designated based on the important person table shown in FIG. Here, 2006 is designated. The search time condition is input by the user.
The result display unit 52 of the output control unit 50 selects the thread ID “116” whose start date and time is 2006 among the thread IDs “116” and “120” included in the important person table (FIG. 17). For the thread ID “116”, “userA”, “userB”, and “userE” are displayed together with the number of points as the important person in descending order of the number of points. Also, from the person correlation table (FIG. 17), in correspondence with the thread ID, a transmission / reception set including “userA” having the highest number of points as an important person and the number of points are read. The user “userB”, “userC”, “userD”, and “userE” are displayed around the “userA” with the highest number of points, and “userA” and each user. The number of points is displayed between “userB”, “userC”, “userD”, and “userE”, and a line with a thickness corresponding to the number of points is drawn. The transmission / reception group to be displayed may be a predetermined number from a group having a large number of points, or a transmission / reception group having a predetermined number of points or more may be displayed.

  FIG. 20 is a diagram showing an overall image of the search screen. In the figure, a field for entering a search word, a thread arranged in chronological order and its important persons, a person correlation diagram, a list of users who appeared as senders or receivers in each thread, and the number of points are displayed. ing.

  In the above, the threads within the period for which the start date / time is specified as the search condition are output, but the threads whose end date / time is within the period specified as the search condition may be output. May be combined. When the end date / time is set as the search condition, the end date / time is read instead of the start date / time in step S240 of FIG. 10, and the search result reading unit 51 reads the end date / time of each processing target thread ID in step S510 of FIG. Include.

  When displaying the user, the user identification information (email address, etc.) read from the sender and recipient fields of the mail may be displayed and stored in the storage means in association with the user identification information. You may read and display information such as your name and department.

[Second Embodiment]
A second embodiment of the present invention will be described. In the first embodiment described above, the mail divided into threads is plotted and displayed in the time series of the start date and time. However, when a large number of threads start at the same time, or the search results There are cases where many threads are applicable. In such a case, it becomes difficult for the searcher to select which thread deals with the content he wants to know. Therefore, in the second embodiment, in order to assist a searcher to easily select a thread that deals with the content that the searcher wants to know, the distribution of the amount of mail over time is visualized and displayed. Enable thread selection that takes into account ups and downs. This allows you to select threads where many emails are being exchanged for the content that the searcher wants to know, or to select threads based on when the buzz was raised, that is, when the amount of email was high It becomes like this.

The difference between the processing of the second embodiment and the first embodiment will be described.
In step S240 of the thread search processing flow shown in FIG. 10, the search unit 42 of the search processing unit 40 searches the mail table stored in the analysis result storage unit 30 using the selected thread ID as a key, and detects the detected record. In step S250, information that associates the selected thread ID with the read date and time is stored. This information is read by the search result reading unit 51 of the output control unit 50 in step S510 of the result display processing flow shown in FIG. The result display unit 52 of the output control unit 50 counts the number of e-mails for each predetermined unit period (quarterly unit, one month unit, etc.) for each thread ID based on the date information corresponding to the thread ID. For each thread, a circle with a size corresponding to the number of mails is displayed for each unit period, or a transition of the number of mails is displayed in a graph. The unit period may be information stored in advance in a storage unit (not shown) of the output control unit 50, or may be set in a search request received by the query receiving unit 41 of the search processing unit 40. Good.

  FIG. 21 is a thread output image when the number of mails is displayed according to the size of the circle. In the figure, for each thread, a circle of a size corresponding to the number of mails is displayed every quarter. As a result, it is visually determined that threads with larger circles displayed than other threads are threads that have become more topical, and that the time when larger circles are displayed is the time when more topics have been raised. It becomes possible to grasp. Here, a circle with a size corresponding to the number of emails is displayed, but the shape of the email may be a shape other than a circle, and red if the amount of email is large, You may display by a color, such as displaying by the gradation which was made into blue.

  FIG. 22 is a thread output image when the number of mails is displayed by a graph. In the figure, the transition of the number of mails in each quarter is displayed in a graph for each thread. In the displayed graph, if you want to see the transition of the number of emails in more detail, specify the display location with the mouse and enter the thread and timing information to be displayed in detail. The result display unit 52 of the output control unit 50 refers to the information of the date corresponding to the thread, and for each thread and time of the detailed display target, each unit period shorter than the currently displayed (one day unit, one week unit) Etc.) is counted and displayed in a graph.

[Third Embodiment]
Next, a third embodiment of the present invention will be described. In the above-described embodiment, when the person correlation table and the important person table are created, one point is uniformly given to the user who is the sender or receiver of the mail. However, for example, an administrator or the like is often set as a broadcaster (CC) as an e-mail address so that information can be shared. Often. Therefore, depending on whether the administrator or the person in charge or the person to be executed is extracted as the key person, the number of points granted to the broadcaster is increased or set to the sender. Increase the number of points granted and change the points added. In other words, it is possible to select and present a user with a high total number of points as an administrator regarding the content to be searched by weighting the number of points given when set to the broadcaster. By weighting the number of points to be given when set to a person, it is possible to select and present a user with a high total number of points as an information sender such as a person in charge or a practitioner regarding the content to be searched Become.

Regarding the processing of the third embodiment, the difference between the first embodiment and the second embodiment will be described.
FIG. 23 is a diagram showing the configuration of the mail table according to the present embodiment. In the figure, the mail table is composed of a plurality of records consisting of mail ID, thread ID, message ID, subject, sender, receiver, broadcaster, date, and body data, and the mail shown in FIG. Broadcaster data is added to the table record. The broadcaster corresponds to the information set in the CC field of the mail.
Then, in step S120 of the thread search processing flow of the first embodiment shown in FIG. 5, the analysis result output unit 23 of the mail collection unit 20 includes the data item of the broadcaster in the record added to the mail table. Write the setting contents of the CC field of the read mail.

  In step S320 of the person correlation table creation processing flow shown in FIG. 11, the search unit 42 of the search processing unit 40 performs the following in order for each record read in step S310. That is, a user set set as a sender-recipient and a user set set as a sender-broadcaster are extracted and set as a user set as a sender and a receiver. One point is added to the user's manager search correspondence point, and three points are added to the user's manager search correspondence point set as a broadcaster. Also, 3 points are added to the information sender search correspondence point of the user set as the receiver, and the information sender search correspondence point of the user set as the sender and the user set as the broadcaster is added. Add 1 point.

  As described above, when extraction of a set of users and assignment of an administrator search correspondence point and an information sender search correspondence point are executed for all processing target thread IDs, in step S340, the search unit 42 selects a thread ID. A list showing a sender-recipient user pair, a sender-broadcaster user pair, and an administrator search correspondence point and an information sender search correspondence point of each user included in these pairs. A person correlation table is generated and stored.

  In step S580 of the result display processing flow shown in FIG. 13, the result display unit 52 of the output control unit 50 displays the user set extracted for each thread. In addition, at this time, while displaying a matrix-like relation diagram showing the correlation of a set of users among a plurality of users, the user having the highest number of search points for administrators, or a predetermined number of users in descending order Information for identifying the user, the user with the highest number of information sender search corresponding points, or information for identifying a predetermined number of users in the descending order is also displayed. Further, the result display unit 52 displays information about which form of the sender-receiver relationship or the sender-broadcaster relationship is large for each user pair.

  FIG. 24 is a display image of a person correlation diagram according to the present embodiment. Here, user A having the highest number of administrator search points is displayed in red, and user B having the highest number of information sender search corresponding points is displayed in blue. In addition, when displaying a predetermined number of users in descending order of the number of points, display in a gradation that the color becomes lighter as the number of points decreases, or search points for managers, information sender search corresponding points It may be displayed in association with the user. The display color can be arbitrarily selected. Moreover, although the group of users is indicated by an arrow, the user at the beginning of the arrow is a sender, and the user at the end of the arrow is a receiver or a broadcaster.

  In step S420 of the important person table creation processing flow shown in FIG. 12, the search unit 42 of the search processing unit 40 extracts users set as a sender, a receiver, or a broadcaster, and transmits the sender or the receiver. If it is set as a receiver, 1 point is added to the administrator search corresponding point if it is set as a broadcaster, and if it is set as a receiver, 3 points are added to the sender or If it is set as a reporter, 1 point may be added to the information sender search corresponding point. Then, in step S540 of the result display processing flow of FIG. 13, the result display unit 52 of the output control unit 50 sets the information sender as a predetermined number of users as managers in the order of descending points corresponding to the manager search for each thread. A predetermined number of users are displayed as information transmission users in order of the number of search corresponding points.

  In the above, both the manager search and the information sender search are performed, but either one may be performed. Moreover, the addition point mentioned above is an example, and can be set suitably.

[Fourth Embodiment]
When sending mail using a mailing list, the recipient (To) is the same. Therefore, in the fourth embodiment, points are added based only on the sender (From) information. Hereinafter, differences from the third embodiment will be described.

  First, a list of mailing addresses (mlist@xxx.jp, members@xxx.jp, etc.) used for the mailing list is stored in advance in a storage unit (not shown) of the mail collection unit 20. Also, as shown in FIG. 25, mailing list flag data is added to the mail table record.

  In step S120 of the thread search processing flow shown in FIG. 5, the analysis result output unit 23 of the mail collection unit 20 includes the setting contents of the read To field in the mailing address held in the storage unit. Judgment is made. When the To field of the read mail is included in the mailing list address, a value indicating the mailing list is written in the mailing list flag of the record added to the mail table.

  In step S320 of the person correlation table creation processing flow shown in FIG. 11, the search unit 42 of the search processing unit 40 performs the following in order for each record read in step S310. That is, a mailing list is extracted in the mailing list flag of the record from which the set of users set as a sender-recipient and the set of users set as a sender-broadcaster are extracted. In the case where the value indicating is not set, the administrator search corresponding point and the information sender search corresponding point are set to each of the sender, the receiver, and the user set as the broadcaster as in the third embodiment. If a value indicating that the mailing list is set is added, the administrator search corresponding point and the information sender search corresponding point are added only to the user set as the sender.

  Further, in step S420 of the important person table creation processing flow shown in FIG. 12, the search unit 42, if a value indicating a mailing list is not set in the mailing list flag of the extracted record that is the origin, , Extract users set as recipients or broadcasters, and if a value indicating mailing list is set, only users set as senders are extracted and each user The extracted number is set as the number of points of the user. Then, in step S540 of the result display process flow of FIG. 13, the result display unit 52 of the output control unit 50 displays a predetermined number of users as important persons in the order of increasing points for each thread.

[Fifth Embodiment]
The difference between the fifth embodiment and the above-described embodiment will be described.
explain. In the above embodiment, the result of morphological analysis of the mail subject (Subject field) is stored in the keyword table. This is because it is possible to extract keywords correctly in the ideal case where the subject of the email matches the content of the text, but when replying, completely different content may be written in the text. is there. For example, for emails with “Commercial Service Deployment” set in the subject line, the main topic is about the next in-house system as the reply is repeated, even if the text on the topic related to commercial service deployment is initially written. The body will be described. Therefore, in the present embodiment, the subject of morphological analysis for extracting keywords is the subject (subject field) and body of the mail.

  Specifically, in step S180 of the analysis processing flow shown in FIG. 5, the mail analysis unit 22 of the mail collection unit 20 performs a morphological analysis on the setting contents of the subject field and body of the mail, and uses the result of the morphological analysis as a keyword. And Then, the analysis result output unit 23 of the mail collection unit 20 adds a record including a record ID, a thread ID, and a keyword to the keyword table stored in the analysis result storage unit 30.

  However, when the morphological analysis is performed on the text as keywords as described above, the number of keywords registered in the keyword table increases because the analysis target is expanded. Therefore, the following keyword filtering process is performed. That is, from the result of morphological analysis on the Subject field and the text, leave only the appropriate part-of-speech character string as a keyword, for example, a general noun / verb as a keyword candidate. Delete inappropriate character strings and register the remaining character strings as keywords. The character string to be deleted is, for example, the words “below”, “get”, “convenient”, “a little”, etc. that can be used in any topic. Store in advance. By performing such keyword filtering processing, only more important and characteristic keywords can be stored in the keyword table.

[Sixth Embodiment]
FIG. 26 is a functional block diagram of the search support apparatus 1a according to the sixth embodiment of the present invention, in which only functional blocks related to the present invention are extracted and shown. In this figure, the same parts as those in the search support apparatus 1 according to the first embodiment shown in FIG. The search support apparatus 1a shown in FIG. 26 is different from the search support apparatus 1 according to the first embodiment shown in FIG. The search processing unit 40 further includes a person information reading unit 44. Here, it is assumed that the output control unit 50 is a searcher's computer terminal, and further includes a search document reception unit 53 for receiving an input of a search word from an input device such as a keyboard.
Hereinafter, differences from the first embodiment will be described.

  The person information generation unit 24 of the mail collection unit 20 analyzes the email data read by the mail reading unit 21, extracts the correspondence between the mail address and the person's name at the mail address, and stores the person information. The data is written in the person table held by the unit 60. The person information reading unit 44 of the search processing unit 40 reads the name information corresponding to the mail address from the person table held in the person information storage unit 60. The person information storage unit 60 can be realized by a database server or the like.

Next, each data held by the search support apparatus 1a according to the present embodiment will be described.
The mail table, thread table, and keyword table stored in the analysis result storage unit 30 according to the present embodiment are the same as the mail table, thread table, and keyword table of the first embodiment shown in FIGS. is there. In this embodiment, the sender and receiver data items of each record in the mail table include the sender and recipient identification information as the sender and receiver identification information in the From field and To field. A case where an email address is set will be described as an example.

  FIG. 27 is a diagram illustrating a configuration of a person table stored in the person information storage unit 60. In the figure, the person table is composed of a plurality of records including data of mail addresses and person names corresponding to the mail addresses. The mail address set in the From field of the mail is stored in the mail address of the person table, and the person name extracted from the body of the mail by the person information generating unit 24 is set in the person name of the person table. .

Next, a processing flow in the search support apparatus 1a will be described.
The analysis processing flow in the mail collection unit 20 is the same as the analysis processing flow in the mail collection unit 20 shown in FIG.

FIG. 28 is a person information generation process flow of the person information generation unit 24.
First, the mail reading unit 21 of the mail collection unit 20 reads a mail that has not yet been subjected to the personal information generation process according to the personal information generation process flow from the mail server as the mail storage unit 10 (step S810). The person information generation unit 24 reads the content from the head to the predetermined line, for example, the fifth line from the mail body field of the mail read in step S810 (step S820). Each line is separated by a line feed symbol. The person information generation unit 24 searches for a line described as “to is” in the five lines read out in step S820 by using a regular expression, and from the line including “is” to immediately before “is”. A sentence is extracted (step S830). At this time, in order to avoid misextraction such as "Thank you for your work" and "I am delighted", I will extract lines that introduce myself in the email, such as "I am Ito". Do not extract lines that use the expression "is." In addition to “~”, a phrase used for the expression of self-introduction such as “I say ~” may be used as a keyword, and a sentence may be cut out from a line including the keyword.

  Next, the person information generation unit 24 extracts words that appear before or after the blank character, punctuation mark, and symbol as delimiters (step S840). As a result, when a name is present following the part where the organization name or greeting is stated, an unnecessary part can be removed. For example, in the case of the line “Ide is in charge of development”, “development Ito” is extracted in step S830, and the word “Ito” after the space is extracted with the space as a delimiter. In the case of the line “Ito @ development in charge”, “development in charge @ Ito” is extracted in step S830, and the word “Ito” before the symbol is extracted with the symbol (@) as a delimiter. Alternatively, in the case of the line “I am in charge of development”, “development in charge, Ito” is extracted in step S830, and “Ito” after the punctuation mark is extracted with the punctuation mark (,) as a delimiter. Note that for a line without a delimiter, the document extracted in step S830 is extracted as it is.

  Subsequently, the person information generation unit 24 performs morphological analysis on the document extracted in step S840, and selects a phrase (A) classified as part of speech as a proper noun (step S850). In this way, by making a rule that a part of speech is a proper noun is a candidate for a person name, noise that is a word other than a name is removed. When the person information generation unit 24 acquires the sender's mail address (B) from the From field of the mail header (step S860), the person / phrase (A) selected in step S850 and the mail address (B) acquired in step S860 are obtained. ) Is temporarily stored in a storage unit (not shown) included in the person information generation unit 24 (step S870).

  Subsequently, the mail reading unit 21 determines whether there is a mail that has not yet been subjected to the personal information generation process in the mail server as the mail storage unit 10 (step S880). If there is an e-mail that has not yet undergone the personal information generation process (step S880: YES), the process returns to step S810 and the above-described process is repeated. Then, when there is no mail in the mail storage unit 10 that has not undergone personal information generation processing (step S880: NO), the person table registration is performed from the combination of the mail address (B) and the phrase (A) stored in step S870. Extract the target set. Specifically, when only a set of the same word (A) is stored for a certain mail address (B), the set of the mail address (B) and the word (A) is set as a person table registration target. Extract as a pair. On the other hand, when a combination with a different phrase (A) is stored for the same email address (B), the most frequently occurring phrase (A) is the name of the sender corresponding to the email address (B). Is extracted as a person table registration target set. For example, three pairs of email address “user@srv.jp” and the phrase “Ito”, one pair of email address “user@srv.jp” and the phrase “Japan”, and email address “user @ srv” .jp ”and the phrase“ Kansai ”are stored, and the phrase“ Ito ”with the highest appearance rate is a phrase indicating the name of the sender of the email address“ user@srv.jp ” Judgment is made, and a set of the mail address “user@srv.jp” and the phrase “Ito” is extracted. The person information generation unit 24 adds a record in which the set of the extracted mail address (B) and the phrase (A) is set to the person table stored in the person information storage unit 60 (step S890).

The analysis processing flow (FIG. 5) and person information generation processing flow (FIG. 28) according to the present embodiment will be described using specific data examples.
29 shows an example of each mail read from the mail storage unit 10, FIG. 30 shows an example of a mail table generated from the mail shown in FIG. 29, and FIG. 31 shows a thread generated from the mail shown in FIG. FIG. 32 shows an example of a table, FIG. 32 shows an example of a keyword table generated from the mail shown in FIG. 29, and FIG. 33 shows an example of a person table generated from the mail shown in FIG.

First, the setting process of the mail table (FIG. 30), thread table (FIG. 31), and keyword table (FIG. 32) when the mail M1a (FIG. 29) is read is the same as the mail M1 of the first embodiment. Since the processing is the same as that for setting the mail table (FIG. 7), thread table (FIG. 8), and keyword table (FIG. 9) from FIG.
After setting the mail table, thread table, and keyword table, the first five lines set in the body of the mail M1a, the first line “Bessho”, the second line “Abe of the development department.”, 3 lines The item “Thank you for your attention”, the fourth line “in the contents of the file sent yesterday”, the fifth line “Because there was some error” (FIG. 28, step S820). Then, the sentence “Abe of the development department” immediately before “is” is extracted from the second line, which is a line describing “to be” (FIG. 28, step S830). Since the extracted document does not contain delimiters, morphological analysis is performed as it is, and “development department” and “Abe” are extracted, and the proper noun “Abe” is selected as a candidate for the person name. (FIG. 28, steps S840 and S850). Then, From “userA@svr.jp” acquired from the mail M1a (FIG. 28, step S860) and “Abe” selected as the candidate for the person name are associated with each other and temporarily stored (FIG. 28). Step S870).

Subsequently, the setting process of the mail table (FIG. 30), thread table (FIG. 31), and keyword table (FIG. 32) when the mail M2a (FIG. 29) is read is the same as the mail M2 of the first embodiment. Since the processing is the same as that for setting the mail table (FIG. 7), thread table (FIG. 8), and keyword table (FIG. 9) from FIG.
After setting the mail table, thread table, and keyword table, the first 5 lines set in the body of the mail M2a are acquired, but here the body is limited to 4 lines, so that 4 lines are acquired. (FIG. 28, step S820). Then, except for the line that contains "~ sama is", the text "Between the system part" is extracted from the first line described as "~ is" immediately before "is" (Fig. 28, step) S830). Since the extracted document does not contain a delimiter, morphological analysis is performed as it is to extract “system part” and “Bessho”, and the proper noun “Bessho” is selected as a candidate for a person name. (FIG. 28, step S840, S850). Then, From “userB@svr.jp” acquired from the mail M2a (FIG. 28, step S860) and “Bessho” selected as a candidate for the person name are associated with each other and temporarily stored (FIG. 28). Step S870).

  The result of performing the processing of FIG. 28 on the mails M1 to M8 shown in FIG. 29 is the mail table shown in FIG. 30, the thread table shown in FIG. 31, and the keyword table shown in FIG. The result of performing the process of step S890 of FIG. 28 from the set of the stored mail address and the selected person name is the person table shown in FIG.

Next, processing when a search is requested will be described.
The processing flow after the query receiving unit 41 of the search processing unit 40 reads the search word received by the result display unit 50 is the same as in FIGS. 10 to 12 of the first embodiment.

  FIG. 34 shows a result display processing flow of the output control unit 50. In the figure, the processes from step S910 to S980 are the same as steps S510 to S580 of FIG. 13 of the first embodiment. However, in step S940, the result display unit 52 extracts, for each thread, a predetermined number of user e-mail addresses as important person e-mail addresses in descending order of the number of user e-mail addresses indicated in the important person table. To do. Then, the mail address of the important person of each extracted thread is displayed together with the number of points corresponding to the mail address of the important person. In step S980, the result display unit 52 extracts, for each thread, a predetermined number of pairs in the descending order of the number of points from the set of the sender email address and the recipient email address shown in the person correlation table. Then, the set of the sender mail address and the receiver mail address extracted for each thread is displayed together with the number of points corresponding to the number of points of the pair of the sender mail address and the receiver mail address sender / receiver.

  Subsequently, when it is set that display of the person name is necessary (step S990: YES), the search result reading unit 51 and the sender mail address of the person correlation table read as the search result in step S910 and The mail address set in the recipient mail address group, the user mail address set in the important person table, or the mail address currently displayed on the screen is output to the search processing unit 40, and the person name Request information.

  When the person information reading unit 44 of the search processing unit 40 reads information on the person name corresponding to the received mail address from the person table stored in the person information storage unit 60, the person information reading unit 44 associates the mail address with the person name. The person name search result is returned to the search processing unit 40. When the search result reading unit 51 receives the person name search result output from the person information reading unit 44, the result display unit 52 correlates with the display of the mail address, and the mail address indicated by the person name search result. The person name associated with is displayed (step S1000).

  Note that in step S910, in accordance with the search result, the mail address set in the combination of the sender mail address and the receiver mail address of the person correlation table included in the search result, the user set in the important person table A person name search result indicating a person name associated with the e-mail address may be received from the search result processing unit 40 in advance.

The processing of the search processing unit 40 and the processing of the output control unit 50 described above will be described using specific data examples.
35 shows an example of a mail table, FIG. 36 shows an example of a thread table, FIG. 37 shows an example of a keyword table, and FIG. 38 shows an example of a person table, before the query receiving unit 41 reads a search keyword. It is assumed that it is set to. Further, it is assumed that “sales” is input as a search keyword. FIG. 39 shows the person correlation table and the important person table created at this time.

  In the thread search process, when a search keyword is read (FIG. 10, step S210), thread IDs set in the keyword table (FIG. 37) are sequentially selected (FIG. 10, step S220). When the thread ID “116” is selected, a record (record ID “340”) in which the search word “sales” is set as a keyword from a record in the keyword table (FIG. 37) in which the thread ID is set. It is detected (FIG. 10, step S230). The start date and time “2006/7/1 15:12” corresponding to the thread ID is read and stored in association with the thread ID “116” (FIG. 10, steps S240 and S250).

  Next, when the thread ID “120” is selected, a record (record ID “372”) in which the search word “sales” is set as a keyword from a record of the keyword table (FIG. 37) in which the thread ID is set. ") Is detected (FIG. 10, step S230). Then, the start date and time “2007/1/7 13:10” corresponding to the thread ID is read and stored in association with the thread ID “120” (FIG. 10, steps S240 and S250).

Next, when the thread ID “144” is selected, it is detected that there is no record in which the search word “sales” is set as a keyword in the keyword table (FIG. 37) in which the thread ID is set. (FIG. 10, step S230).
As a result, the processing target thread IDs “116” and “120” and the start dates and times “2006/7/1 15:12” and “2007/1/7 13:10” are stored (FIG. 10, step). S250).

Subsequently, a person correlation table creation process (FIG. 11) is started, and process target thread IDs are sequentially selected. When the thread ID “116” is selected, the records (mail IDs “500” to “506”) of the mail table (FIG. 35) in which the thread ID is set are specified (FIG. 11, step S310). As shown in FIG. 39, “userA@svr.jp and userB@svr.jp”, “userA@svr.jp and userC @ svr” are obtained from information of the sender mail address and the receiver mail address of the specified record. .jp "," userA@svr.jp and userD@svr.jp "," userA@svr.jp and userE@svr.jp ", and" userB@svr.jp and userE@svr.jp " The numbers extracted from these sets are “4”, “2”, “2”, “2”, and “3”, respectively (FIG. 11, step S320).
Next, when the thread ID “120” is selected, the records (mail IDs “507” to “511”) of the mail table (FIG. 35) in which the thread ID is set are specified (FIG. 11, step S310). ). As shown in FIG. 39, “userA@svr.jp and userB@svr.jp”, “userA@svr.jp and userC @ svr” are obtained from information of the sender mail address and the receiver mail address of the specified record. .jp ”,“ userA@svr.jp and userD@svr.jp ”,“ userA@svr.jp and userE@svr.jp ”are extracted, and the number of these extracted groups is“ 2 ”. , “2”, “3”, “1” (FIG. 11, step S320).
As a result, for each thread ID, a person correlation table (FIG. 39) is generated that shows a set of the extracted mail addresses of the users and points corresponding to the number of extractions of the set (FIG. 11, step). S340). Here, the extraction number 1 is 1 point.

Subsequently, the important person table creation process (FIG. 12) is started, and the process target thread IDs are sequentially selected. When the thread ID “116” is selected, the records (mail IDs “500” to “506”) of the mail table (FIG. 35) in which the thread ID is set are specified (FIG. 12, step S410). As shown in FIG. 39, as user email addresses set in the sender email address or recipient email address of this identified record, “userA@svr.jp”, “userB@svr.jp”, “ “userC@svr.jp”, “userD@svr.jp”, “userE@svr.jp” are extracted, and the number of appearances of the extracted user's mail address is “10”, “7”, “2”, “2” ”And“ 5 ”are measured.
Next, when the thread ID “120” is selected, the records (mail IDs “507” to “511”) of the mail table (FIG. 35) in which the thread ID is set are specified (FIG. 12, step). S420). As shown in FIG. 39, as user email addresses set in the sender email address or recipient email address of this identified record, “userA@svr.jp”, “userB@svr.jp”, “ “userC@svr.jp”, “userD@svr.jp”, “userE@svr.jp” are extracted, and the number of appearances of the extracted user's email address is “8”, “2”, “2”, “3” "," 1 "is measured.
Thereby, for each thread ID, an important person table indicating these extracted user mail addresses and points corresponding to the number of appearances is generated (FIG. 12, step S440). Here, the number of appearances is 1 point.

FIG. 40 is an important person output image based on the important person table shown in FIG. 39, and shows an image in which threads are displayed in time series. Here, there are three important persons in descending order of the number of points. Note that the number of points may be equal to or greater than a predetermined number, and the condition is stored in advance in the output control unit 50 or input by the user.
In FIG. 40, the thread IDs “116” and “120” included in the important person table (FIG. 39) are assigned the start dates and times “2006/07/01 15:12” and “2007/07/08 23:11”. On the time axis corresponding to “”. For the thread ID “116”, “userA@svr.jp”, “userB@svr.jp”, “userE@svr.jp” are displayed in descending order of the number of points, and for the thread ID “120”, “UserA@svr.jp”, “userD@svr.jp”, “userB@svr.jp”, and “userC@svr.jp” are displayed in descending order of the number of points. Since “userB@svr.jp” and “userC@svr.jp” have the same number of points, both are displayed.

FIG. 41 shows a display image of the important person and the person correlation diagram when the search time is designated based on the important person table shown in FIG. Here, 2006 is designated. The search time condition is input by the user.
The result display unit 52 of the output control unit 50 selects the thread ID “116” whose start date and time is 2006 among the thread IDs “116” and “120” included in the important person table (FIG. 39). For the thread ID “116”, “userA@svr.jp”, “userB@svr.jp”, “userE@svr.jp” are displayed together with the number of points as the important person in descending order of the number of points. . Further, from the person correlation table (FIG. 39), the transmission / reception set including “userA@svr.jp” having the highest number of points as an important person and the number of points corresponding to the thread ID are read. And, centering on “userA@svr.jp” with the highest number of points, the email addresses of each user who will be sending and receiving “userB@svr.jp”, “userC@svr.jp”, “userD@svr.jp” ”And“ userE@svr.jp ”are displayed around it, and“ userA@svr.jp ”and the email addresses of each user“ userB@svr.jp ”,“ userC@svr.jp ”,“ userD @ svr ” The number of points is displayed between “.jp” and “userE@svr.jp”, and a line with a thickness corresponding to the number of points is drawn. It should be noted that the set of mail addresses of the sender / receiver to be displayed may be a predetermined number from a group having a large number of points, or a group of transmissions / receptions having a predetermined point number or more may be displayed.

  FIG. 42 shows an overall image of the search screen. In the figure, the check box for instructing the display of the person name is not checked. At this time, the search screen includes a field for entering a search word, a thread arranged in chronological order and the mail address of the important person, the mail address of the person included in the person correlation chart and the person correlation chart, and the transmission to each thread. A list of e-mail addresses of users who have appeared as subscribers or recipients and the number of points are displayed.

  FIG. 43 is a diagram showing an overall image of the search screen when a check box for instructing display of a person name is checked. At this time, the search screen includes a field for entering a search word, a thread arranged in chronological order and the mail address of the important person, the mail address of the person included in the person correlation chart and the person correlation chart, and the transmission to each thread. A list of e-mail addresses of users who appear as users or recipients and the number of points are displayed, and each e-mail address displays a person name associated with the e-mail address.

  It should be noted that the second to fifth embodiments can be combined with this embodiment.

[Seventh Embodiment]
In the seventh embodiment, the utterance content of each participant in the conference is converted into text by voice recognition and formed into a data format similar to that of mail, thereby applying the same processing as in the above embodiment to search for important persons. It is possible to grasp the person correlation. Hereinafter, the difference between the seventh embodiment and the first embodiment will be described.

  FIG. 44 is a functional block diagram of the search support apparatus 1b according to the seventh embodiment of the present invention, in which only functional blocks related to the present invention are extracted and shown. In this figure, the same parts as those in the search support apparatus 1 according to the first embodiment shown in FIG. The search support device 1b shown in FIG. 44 is different from the search support device 1 according to the first embodiment shown in FIG. 1 in that the voice recognition unit 80 is connected to the mail storage unit 10 and the voice data storage unit 70 is a voice. It is connected to the recognition unit 80.

  The voice data storage unit 70 can be realized by, for example, a database server or the like. The voice data storage unit 70 is a meeting name, voice data obtained by collecting voices during the meeting, and data, meeting date and time, meeting participants and moderators. Memorize the meeting file set with the e-mail address. It is assumed that one meeting file is generated for one meeting.

  FIG. 45 is a block diagram showing a configuration of the voice recognition unit 80. The voice recognition unit 80 can be realized by a computer device such as a server, for example, and includes a speaker data storage unit 81, a voice data division unit 82, a speaker recognition unit 83, a voice text conversion unit 84, and a mail conversion unit. 85. The speaker data storage unit 81 stores voice feature information in which a mail address as user specifying information for specifying each person is associated with a parameter indicating a voice feature of the person.

  The voice data dividing unit 82 reads the conference file from the voice data storage unit 70, and generates divided voice data in which the voice data in the read conference file is divided into speech sections in which one speaker is speaking. The speaker recognition unit 83 specifies a speaker from the divided voice data and the voice feature information of each speaker stored in the speaker data storage unit 81, and reads the mail address of the specified speaker. The voice text conversion unit 84 generates text data in which the utterance content is converted into text from the divided voice data. The e-mail conversion unit 85 sets e-mail data in which the e-mail address of the speaker, the e-mail address of the participant in the conference file, the conference time, and the text data of the utterance content are set in the From field, To field, Subject field, and e-mail body, respectively. Is written into the mail storage unit 10.

Next, processing of the search support apparatus 1b according to the present embodiment will be described.
FIG. 46 shows a mail format data generation processing flow in the voice recognition unit 80 according to the present embodiment. It is assumed that the speech feature information of the conference participant is registered in advance in the speaker data storage unit 81 of the speech recognition unit 80. After the conference ends, the conference file is registered in the audio data storage unit 70.

  For example, the meeting organizer inputs the meeting name, meeting participants and the moderator's e-mail addresses in advance to a PC (personal computer), etc., and inputs the voice collection start instruction to the PC at the start of the meeting. Sometimes input voice collection stop instruction. The PC generates voice data of the voice during the meeting collected by the microphone included in the PC until the voice collection stop instruction is inputted after the voice collection start instruction is inputted. When the voice collection stop instruction is input and the generation of the voice data is finished, the PC finishes generating the conference name, the conference participants and the moderator mail addresses, the generated voice data, and the conference date and time. A conference file in which the generation date and time of the audio data is set is generated and written in the audio data storage unit 70. The PC may be a device that implements the voice data storage unit 70 and the voice recognition unit 80.

  The voice data dividing unit 82 of the voice recognition unit 80 reads from the voice data storage unit 70 a conference file that has not yet been subjected to the mail format data generation process according to the mail format data generation process flow (step S1110). Note that a conference file designated as a processing target by the user may be read. When the voice data dividing unit 82 acquires the voice data set in the read conference file, the voice data dividing unit 82 extracts the divided voice data obtained by dividing the voice data into speech sections in which one speaker is speaking (step S1120). ). At this time, if the first divided audio data is extracted, the divided audio data is extracted from the beginning of the audio data excluding the divided audio data extracted immediately before from the beginning of the audio data if extracted for the second time or later. Extract. Note that one utterance section can be, for example, a silent section to the next silent section, or a section while one speaker is speaking. The voice data dividing unit 82 determines the elapsed time from the voice collection start time corresponding to the divided voice data, that is, how much the start part of the utterance section indicated by the divided voice data is from the beginning in the in-conference voice indicated by the voice data. Get whether the time has passed.

  Next, the speaker recognizing unit 83 uses an existing speaker recognition technique to identify who speaks the voice indicated by the divided voice data extracted in step S1120. For example, it is determined whether the voice (utterance) pattern indicated by the divided voice data matches the voice feature information of any user stored in the speaker data storage unit 81, for example, data indicating the characteristics of the voiceprint. (Step S1130). At this time, only the audio feature information specified by the mail address registered in the conference file read out in step S1110 may be used.

  If there is matching voice feature information (step S1130: YES), the speaker recognition unit 83 acquires the e-mail address set in the matching voice feature information from the speaker data storage unit 81, and this acquisition is performed. It is determined whether the e-mail address of the speaker is the same as the e-mail address of the moderator set in the conference file (step S1140). If the speaker's email address is not the same as the presenter's email address (step S1140: NO), the speaker recognition unit 83 uses the email address read from the matching voice feature information as the speaker's email address. Is temporarily stored in the storage means (step S1150). The speech text conversion unit 84 converts the utterance content into text from the divided speech data extracted in step S1120, using existing speech recognition technology, and generates text data (step S1160).

  The mail conversion unit 85 generates mail data based on the speaker's mail address, the text data of the utterance content, and the divided voice data, and writes the mail data in the mail information storage unit 70 (step S1170). Specifically, first, the mail conversion unit 85 includes the speaker's email address stored in the speaker recognition unit 83 in step S1150, the participant's email address in the conference file read out in step S1110, and the conference file. Of the conference name and the text data of the utterance content generated in step S1160 are generated in the From field, To field, Subject field, and mail text respectively. The To field may be set by excluding the speaker's email address from the participant's email address in the conference file.

  Subsequently, the mail conversion unit 85 generates a unique Message-ID for each utterance, that is, for each divided voice data, and sets it in the Message-ID field of the mail data. Message-ID includes a specific character string to indicate that it is mail data generated based on the utterance at the conference, such as “0001@kaigi20081122.svr.jp”, in this case, “kaigi” Shall. The mail conversion unit 85 sets the Message-ID of other mail data already generated for other divided voice data extracted from the voice data in the same conference file in the References field. Here, since the divided voice data is extracted in order from the beginning of the voice data, the Message-ID of the mail data with the utterance content set at the time point back from the divided voice data currently being processed is acquired and set Will be. Therefore, in the case of the divided audio data extracted first, the References field is not set. Furthermore, the mail conversion unit 85 sets the date and time obtained by adding the elapsed time acquired in step S1120 to the meeting date and time acquired from the conference file in the Date field of the mail data.

  The voice data dividing unit 82 determines whether or not there is a next utterance section in the voice data (step S1180). If there is a next utterance section in the voice data (step S1180: YES), the processing from step S1120 is repeated. When the extraction of the divided audio data up to the end of the audio data is completed (step S1180: NO), the process ends.

If it is determined in step S1130 that there is no matching audio feature information (step S1130: YES), the process of step S1180 is performed. As a case where there is no matching voice feature information, for example, simultaneous speech of a plurality of speakers or a case where the voice feature information of the talker is not registered in the talker data storage unit 81 is assumed.
In step S1140, if the speaker's email address is the same as the moderator's email address registered in the conference file (step S1140: YES), the process of step S1180 is performed. Thus, a moderator who has a lot of remarks but is not an important person related to the topic of the conference is not detected as an important person at the time of search. Note that the process of step S1140 is not performed when it is desired to search for an important person related to the topic of the meeting for the presenter.

  FIG. 47 shows an analysis processing flow in the mail collection unit 20 according to the present embodiment. In the figure, the processing from step S1210 to S1270 is the same as the processing from step S110 to S170 in FIG. 5 in the first embodiment.

  The mail analysis unit 22 determines whether the setting content of the Message-ID field of the read mail includes a character string indicating that the mail is generated based on the conference such as “kaigi” (step S1280). ). When the character string indicating that the mail is generated based on the conference is not included (step S1280: NO), the same processing as steps S180 to S190 of FIG. 5 in the first embodiment is performed (step S1290). ~ S1300). However, in step S1290, the mail analysis unit 22 performs morphological analysis on the setting contents of the mail Subject field and the mail text, and uses the result of the morphological analysis as a keyword.

  On the other hand, when the character string indicating that the mail is generated based on the meeting is included (step S1280: YES), the mail analysis unit 22 performs morphological analysis on the setting contents of the subject field and the text of the mail. The result of the morphological analysis is set as a keyword candidate (step S1310). When the mail analysis unit 22 selects one of the keyword candidates extracted in step S1310 (step S1320), the mail analysis unit 22 stores the thread ID and the selected keyword candidate corresponding to the thread ID. The number of appearances is updated to the number of appearances obtained by adding 1 (step S1330). When the thread ID and the number of appearances corresponding to the selected keyword candidate are not stored, the appearance number “1” is written in association with the thread ID and the selected keyword candidate. The mail analysis unit 22 determines whether or not the updated number of appearances has reached a threshold (for example, 3 times) of the number of appearances for which the keyword candidate should be registered as a keyword (step S1340).

  If it is determined in step S1340 that the number of appearances after the update is less than the threshold or greater than the threshold (step S1340: ≠), the mail analysis unit 22 has not yet selected the keyword candidates extracted in step S1310. It is determined whether there are keyword candidates (step S1360).

  In step S1340, when the number of appearances after the update reaches a threshold value (step S1340: =), the analysis result output unit 23 uses this keyword candidate as a keyword and assigns a new record ID. The analysis result output unit 23 adds a record including the newly assigned record ID, thread ID, and keyword to the keyword table (step S1350). The mail analysis unit 22 determines whether there is a keyword candidate that has not yet been selected among the keyword candidates extracted in step S1310 (step S1360).

  If there is a keyword candidate that has not yet been selected (step S1360: YES), the mail analysis unit 22 repeats the processing from step S1320. If all the extracted keyword candidates are selected (step S1360: NO), the processing from step S1300 is performed.

In the case of a conference, if all the words included in the utterance content are used as keywords, the amount of keywords may be enormous, or content not related to the main topic of the conference may be extracted as keywords. . Therefore, only the word that appears many times during the meeting is selected as a keyword by the above processing. Note that if the keyword is one that has been remarked even once during the meeting, the processing of steps S1280 and S1310 to S1360 need not be performed.
In addition, in order to prevent the amount of keywords registered in the keyword table from increasing or unnecessary keywords from being registered, the fifth embodiment is applied to filter the keywords and keyword candidates. May be performed. Alternatively, only the Subject field may be morphologically analyzed to extract keywords and keyword candidates.

  Next, processing in the search processing unit 40 according to the present embodiment will be described. The thread search processing flow executed when the query receiving unit 41 of the search processing unit 40 of the present embodiment receives and executes a search request is the same as the thread search processing flow according to the first embodiment shown in FIG.

FIG. 48 shows a person correlation table creation process flow according to this embodiment.
The search unit 42 of the search processing unit 40 has not yet performed the person correlation table creation process by the person correlation table creation process flow among the thread IDs stored in step S250 of FIG. 10 (hereinafter, the process target thread ID). Select an ID. The search unit 42 searches the mail table, identifies and reads the record in which the selected thread ID is set (step S1410).

  The search unit 42 extracts a set of users set as a sender and a receiver as a set of senders and receivers from the record read in step S1410, and calculates the number extracted for each of the extracted sets of senders and receivers. Count (step S1420). In addition, even if the sender and the user of the receiver are switched, it is assumed that they are the same set. The search unit 42 determines whether the message ID of the record read in step S1410 includes a character string indicating mail generated based on the utterance in the conference such as “kaigi” (step S1430). .

  When the character string indicating that the mail is generated based on the utterance in the conference is not included (step S1430: NO), the sender / receiver pair is extracted for each sender / receiver pair extracted in step S1420. The number is assigned as a point as it is (step S1440). On the other hand, when the character string which shows that it is the mail produced | generated based on the speech in a meeting is contained (step S1430: YES), the group of the said sender / receiver is set for each sender / receiver group extracted in step S1420. A number obtained by multiplying the extracted number by a predetermined coefficient a such as 0.1 (0 <a <1) is given as a point (step S1450).

  The search unit 42 determines whether there is a thread ID that has not yet undergone the person correlation table creation process in the process target thread ID (step S1460). If there is a thread ID that has not yet been subjected to the person correlation table creation process (step S1460: NO), the process from step S1410 is repeated. When the person correlation table creation processing is executed for all the processing target thread IDs (step S1460: YES), the search unit 42 corresponds to the sender-recipient user pair and the pair for each thread ID. A person correlation table that is a list showing the points that have been created is generated and stored (step S1470).

FIG. 49 shows an important person table creation processing flow according to this embodiment.
The search unit 42 of the search processing unit 40 selects a thread ID that has not yet been subjected to the important person table creation process in the important person table creation process flow from the process target thread IDs stored in step S250 of FIG. . The search unit 42 searches the mail table, identifies and reads the record in which the selected thread ID is set (step S1510). The search unit 42 determines whether or not the message ID of the record read in step S1510 includes a character string indicating mail generated based on the conference such as “kaigi” (step S1520).

  When the character string indicating that the mail is generated based on the conference is not included (step S1520: NO), the search unit 42 is a user set as a sender or a receiver from the read record. The number of appearances, which is the number set as the sender or receiver, is measured, and the measured number is stored as the user's point (step S1530).

  On the other hand, when the character string indicating that the mail is generated based on the meeting is included (step S1520: YES), the search unit 42 is set as the sender or the receiver from the read record. And the number of the extracted users set as the sender and the number set as the receivers are respectively measured (step S1540). For each extracted user, the search unit 42 multiplies the number of the user set as the sender by the predetermined coefficient b and the number of the user set as the receiver by the predetermined coefficient c. The multiplied number is summed to calculate the user's points. Here, the coefficient b and the coefficient c are 0 <b <1, 0 <c <1, and b> c. The coefficient b or the coefficient c is set to the same value as the coefficient a. The search unit 42 stores the extracted points of each user (step S1550).

  If there is a thread ID that has not yet undergone the important person table creation process in the process target thread ID (step S1560: NO), the search unit 42 repeats the process from step S1510. When the important person table creation process is executed for all process target thread IDs (step S1560: YES), the user extracted in step S1530 or S1550 and the points of the user are indicated for each thread ID. An important person table as a list is generated and stored (step S1570).

  The subsequent result display processing according to this embodiment is the same as that of the first embodiment shown in FIG.

  In the case of a conference, since there are many times that a speaker is switched between participants, a large amount of mail indicating the utterance content is generated from one conference file. Therefore, points are given by weighting one normal mail so as to correspond to a plurality of utterances at the conference. In the above, in the case of an email generated from a conference file, the number of pairs of senders and receivers extracted is multiplied by a coefficient a = 0.1, as pointed. It is shown that the weight ratio is 10: 1], that is, one normal mail is weighted corresponding to 10 utterances in the conference. In addition, by setting coefficient b> coefficient c, it is possible to weight participants who speak in the conference from other participants in the conference and regard the speaker as a more important person. In addition, by setting the coefficient b or the coefficient c to the same value as the coefficient a, the ratio of the weights of one normal mail and one utterance at the meeting is made the same in the person correlation table and the important person table. The coefficients a, b, and c can be arbitrarily set within the range of the above-described conditions.

  In step S1540 of the important person table creation processing flow shown in FIG. 49, the search unit 42 extracts a user set as a sender from the read record, and the extracted user is set as a sender. Alternatively, the number obtained by multiplying the number by the coefficient a may be calculated as the user's point.

  In addition, although the difference between this embodiment and 1st Embodiment was demonstrated in the above, it is also possible to combine with 2nd, 4th, and 6th embodiment.

  The search support apparatuses 1, 1a, 1b have a computer system inside. The operation processes of the mail collection unit 20, the search processing unit 40, the output control unit 50, and the voice recognition unit 80 described above are stored in a computer-readable recording medium in the form of a program. The above processing is performed by reading and executing the computer system. Here, the computer system includes a CPU, various memories, an OS, and hardware such as peripheral devices.

Further, the “computer system” includes a homepage providing environment (or display environment) if a WWW system is used.
The “computer-readable recording medium” refers to a storage device such as a flexible medium, a magneto-optical disk, a portable medium such as a ROM and a CD-ROM, and a hard disk incorporated in a computer system. Furthermore, the “computer-readable recording medium” dynamically holds a program for a short time like a communication line when transmitting a program via a network such as the Internet or a communication line such as a telephone line. In this case, a volatile memory in a computer system serving as a server or a client in that case, and a program that holds a program for a certain period of time are also included. The program may be a program for realizing a part of the functions described above, and may be a program capable of realizing the functions described above in combination with a program already recorded in a computer system.

DESCRIPTION OF SYMBOLS 1 ... Search assistance apparatus 10 ... Mail memory | storage part 20 ... Mail collection part 21 ... Mail reading part (mail reading means)
22 ... Mail analysis section (mail analysis means)
23. Analysis result output section (analysis result output means)
24 ... person information generation unit 30 ... analysis result storage unit (analysis result storage means)
40 ... Search processing unit 41 ... Query receiving unit (receiving unit)
42 ... Search part (search means)
43 ... Search result output unit 44 ... Person information reading unit 50 ... Output control unit (output control means)
51 ... Search result reading unit 52 ... Result display unit 53 ... Search document receiving unit 60 ... Person information storage unit (person information storage means)
70: Audio data storage unit (audio data storage means)
80 ... voice recognition unit 81 ... speaker data storage unit (speaker data storage means)
82: Audio data dividing unit (audio data dividing means)
83 ... Speaker recognition unit (speaker recognition means)
84 ... voice text conversion unit (voice text conversion means)
85 ... Mail conversion part (Mail conversion means)

Claims (27)

Message identification information for identifying each electronic mail, thread identification information for identifying the thread to which the electronic mail belongs, a mail table composed of records including at least information on a sender and a receiver, and thread identification information for each thread And an analysis result storage means for storing a keyword table indicating information associating the keywords with each other,
A mail reading means for reading an e-mail from a mail storage means storing the e-mail;
Mail analysis means for acquiring message identification information of other related electronic mail from the information of the header of the electronic mail read by the mail reading means, and determining whether the acquired message identification information is registered in the mail table When,
If it is determined by the mail analysis means that it is not registered, thread identification information of the read e-mail is newly generated, and if it is determined that it is registered, the related other e-mail The thread identification information is acquired from the mail table record specified by the message identification information of the message, the message identification information acquired from the header information of the read electronic mail, the sender and the receiver, and the newly generated thread identification information or A record in which thread identification information acquired from the mail table is set is added to the mail table, and a keyword obtained as a result of morphological analysis of information obtained from the read electronic mail is associated with the thread identification information. Analysis result output means for writing to the keyword table;
Accepting means for receiving search conditions,
The mail table specified by the thread identification information is acquired for each thread identification information obtained from the keyword table, by acquiring thread identification information of a keyword that matches the search word indicated by the search condition received by the receiving means. Search means for acquiring a user set as a sender or a receiver from the record and a number of the user set as a sender or a receiver,
Output control means for outputting a user whose number acquired by the search means is greater than a predetermined condition for each thread identification information;
Equipped with a,
The mail storage means sets a user who is a speaker in a conference as a sender, a participant of the conference as a receiver, a conference name of the conference as a subject, and a utterance content of the user as a speaker as a text. In addition, the message identification information of other e-mails that set other utterance contents in the meeting is set as the message identification information of other related e-mails including information indicating that it was generated based on the meeting. Remember emails,
The information indicating that the analysis result output means is generated based on the conference in the record to be added to the mail table when the information indicating that the analysis result output unit is generated based on the conference is set in the read e-mail. Set further,
The search means specifies, for each acquired thread identification information, each record set in the mail table specified by the thread identification information, and information indicating that the specified record is generated based on a meeting Is set, a predetermined point is given to the user set as the sender in each record, and a point lower than the predetermined point is given to the user set as the receiver. If the information indicating that it has been generated based on the above is not set, a point weighted more than the predetermined point is given to the user set as the sender and the user set as the receiver in each record And
The output control means outputs, for each thread identification information, information indicating a user whose total number of points given by the search means is greater than a predetermined condition.
A search support apparatus characterized by that. The output control means outputs users whose number acquired by the search means is larger than a predetermined condition for each thread identification information in an order corresponding to the number set by the user. The search support apparatus according to 1. The search means acquires thread identification information of a keyword that matches the search word indicated by the search condition received by the reception means from the keyword table, and the acquired thread identification information is identified by the thread identification information. Obtain a set of users set as sender and receiver from the record in the mail table,
The output control means outputs a set of users acquired by the search means for each thread identification information.
The search support apparatus according to claim 1, wherein the search support apparatus is a search support apparatus. The search means acquires thread identification information of a keyword that matches the search word indicated by the search condition received by the reception means from the keyword table, and the acquired thread identification information is identified by the thread identification information. Obtain a set of users set as sender and receiver from the record in the mail table,
The output control unit outputs, for each thread identification information, information indicating a correlation between a plurality of users included in the acquired user set based on the user set acquired by the search unit.
The search support apparatus according to claim 1, wherein the search support apparatus is a search support apparatus. The search means further obtains the number of sets of the user set from the mail table record specified by the thread identification information,
The output control means further outputs information corresponding to the number of sets of users.
The search support apparatus according to claim 3, wherein the search support apparatus is a search support apparatus. The mail table record further includes broadcaster information;
The analysis result output means adds, to the mail table, a record further setting information of the broadcaster acquired from the header information of the read electronic mail,
For each thread identification information acquired, the search means has a predetermined point for a user set as a sender and a user set as a receiver in each record of the mail table specified by the thread identification information. And giving a point weighted by a predetermined weight from the predetermined point to a user set as a broadcaster,
The output control means outputs, for each thread identification information, information indicating a user whose total number of points given by the search means is greater than a predetermined condition.
Search support device according to any one of claims 1 to 5, characterized in that. For each acquired thread identification information, the search means gives a predetermined point to a user set as a receiver in each record of the mail table specified by the thread identification information, and sets as a sender A point given a predetermined weight than the predetermined point is given to the user who has been,
The output control means outputs, for each thread identification information, information indicating a user whose total number of points given by the search means is greater than a predetermined condition.
Search support device according to any one of claims 1 to 5, characterized in that. For each acquired thread identification information, the search means is set as a sender when the receiver set in each record of the mail table specified by the thread identification information is an address of a mailing list. Given points to users who are not mailing list addresses, give given points to users set as senders and users set as recipients,
The output control means outputs a user whose total number of points given by the search means is greater than a predetermined condition for each thread identification information.
Search support device according to any one of claims 1 to 5, characterized in that. When the information indicating that the data is generated based on the audio of the meeting is set in each specified record, the search unit only applies the predetermined information to a user set as a sender in each record. The search support apparatus according to any one of claims 1 to 8, wherein the points are given. The search means acquires thread identification information of a keyword that matches the search word indicated by the search condition received by the reception means from the keyword table, and the acquired thread identification information is identified by the thread identification information. Information indicating that a set of users set as a sender and a receiver and the number of sets of the set of users are obtained from the record in the mail table and generated based on a meeting in the record. When set, information indicating that a predetermined point is assigned to each user group according to the number of the user group set and information indicating that the record is generated based on the conference is set in the record. If the user group is not set, the user group is set with points weighted from the predetermined points. Granted according to the number,
The output control means outputs information corresponding to each user set point,
The search support apparatus according to any one of claims 1 to 9, wherein When the information indicating that the analysis result output means is generated based on the conference is set in the read e-mail, the result obtained by morphological analysis of the information obtained from the read e-mail is set as the keyword candidate, When the keyword candidate is obtained a predetermined number of times or more from the email corresponding to the thread identification information, the keyword candidate is associated with the thread identification information and written to the keyword table.
Search support device according to any one of claims 1 to 10, characterized in that. Audio data storage means for storing conference data including conference name, audio data of audio during the conference, and participants indicating users who participated in the conference;
Speaker data storage means for storing voice feature information which is a parameter indicating the voice feature of each user;
A voice data dividing unit that reads a conference file from the voice data storage unit, and generates divided voice data in which the voice data in the read conference file is divided into speech sections in which one user is speaking;
Speaker recognition means for identifying a user who is a speaker from the divided voice data and voice feature information of each user stored in the speaker data storage means;
Voice text conversion means for generating text data obtained by converting the speech content into text from the divided voice data;
The speaker specified by the speaker recognition means is the sender, the participant in the conference data is the receiver, the meeting name in the conference data is the subject, and the text data generated by the speech text conversion means is the body text. To the other divided voice data extracted from the same voice data as newly generated message identification information, information indicating that it was generated based on the meeting, and message identification information of other related emails A mail conversion unit that generates an email in which the message identification information of the email generated based on the email is set, and writes the email in the email storage unit;
Search support device according to any one of claims 1 to 11, characterized by further comprising a. The conference data further includes information on the date and time of the conference.
The audio data dividing means acquires an elapsed time from the beginning of the sound indicated by the audio data corresponding to the audio indicated by the divided audio data,
The mail conversion means generates an e-mail that further sets a date and time obtained by adding an elapsed time acquired by the voice data dividing means to a meeting date and time in the meeting data.
The search support apparatus according to claim 12 , wherein:   Message identification information for identifying each electronic mail, thread identification information for identifying the thread to which the electronic mail belongs, a mail table composed of records including at least information on a sender and a receiver, and thread identification information for each thread And an analysis result storage means for storing a keyword table indicating information associating the keywords with each other,
  A mail reading means for reading an e-mail from a mail storage means storing the e-mail;
  Mail analysis means for acquiring message identification information of other related electronic mail from the information of the header of the electronic mail read by the mail reading means, and determining whether the acquired message identification information is registered in the mail table When,
  If it is determined by the mail analysis means that it is not registered, thread identification information of the read e-mail is newly generated, and if it is determined that it is registered, the related other e-mail The thread identification information is acquired from the mail table record specified by the message identification information of the message, the message identification information acquired from the header information of the read electronic mail, the sender and the receiver, and the newly generated thread identification information or A record in which thread identification information acquired from the mail table is set is added to the mail table, and a keyword obtained as a result of morphological analysis of information obtained from the read electronic mail is associated with the thread identification information. Analysis result output means for writing to the keyword table;
  Accepting means for receiving search conditions,
  The mail table specified by the thread identification information is acquired for each thread identification information obtained from the keyword table, by acquiring thread identification information of a keyword that matches the search word indicated by the search condition received by the receiving means. Search means for acquiring a user set as a sender or a receiver from the record and a number of the user set as a sender or a receiver,
  Output control means for outputting a user whose number acquired by the search means is greater than a predetermined condition for each thread identification information;
  With
  The mail table record further includes broadcaster information;
  The analysis result output means adds, to the mail table, a record further setting information of the broadcaster acquired from the header information of the read electronic mail,
  For each acquired thread identification information, the search means performs predetermined management for a user set as a sender and a user set as a receiver in each record of the mail table specified by the thread identification information. And assigning an administrator search support point higher than the predetermined administrator search support point to the user set as a broadcaster,
  The output control means outputs, for each thread identification information, information indicating, as an administrator, a user whose total number of administrator search corresponding points is greater than a predetermined condition.
  A search support apparatus characterized by that. The analysis result storage means further stores thread identification information and a thread table indicating information on a start date and time of the thread specified by the thread identification information,
The analysis result output means associates the thread identification information with the start date and time indicating the date and time information obtained from the header information of the email when the thread identification information of the read email is newly generated. Write to the thread table
The search means further reads out information of a start date corresponding to the thread identification information acquired from the keyword table,
The output control means outputs the user or the set of users for thread identification information of a start date and time that matches the period condition received.
Search support device according to any one of claims 14 claim 1, wherein the. The mail table record further includes date information;
The analysis result output means adds a record further setting date information acquired from the header information of the read e-mail to the mail table,
The search means further acquires date information from the record of the mail table specified by the acquired thread identification information,
The output control means outputs, for each thread identification information, information indicating the number of e-mails for each predetermined period based on the date information corresponding to each thread identification information acquired by the search means.
Search support device according to any one of claims 1 to 15, characterized in that. 17. The analysis result output means writes a subject obtained from the read electronic mail or a result of morphological analysis of subject and body information as a keyword in the keyword table. The search support device according to any one of the above. The search support apparatus according to claim 17, wherein the analysis result output means writes a result of morphological analysis with a predetermined part of speech as a keyword in the keyword table. The analysis result output means writes a result obtained by removing a predetermined unnecessary character string from the results of morphological analysis as a keyword in the keyword table. Search support device. Person information storage means for storing a person table indicating information that associates information for identifying a user with the name of the user;
From the information in the body of the email read by the mail reading means, extract a line containing a predetermined word used for self-introduction, and a proper noun among words obtained by morphological analysis of the extracted line A personal information generating unit that associates the selected phrase with information identifying the sender acquired from the read header information of the e-mail and writes it in the person table;
The output control means outputs the name of the user read from the person table corresponding to the information specifying the user to be output or the information specifying the users constituting the set of users to be output To
Search support device according to any one of claims 19 claim 1, wherein the. The person information generating means extracts a line including a predetermined word and phrase used for self-introduction from a line from the beginning to a predetermined number of lines set in the body information of the e-mail. The search support device according to claim 20. The person information generating means extracts a phrase before or after a delimiter from a line extracted as including a predetermined phrase used for self-introduction, performs morphological analysis, and obtains a phrase obtained as a proper noun and The search support apparatus according to claim 20 or 21, wherein the information identifying the sender acquired from the read header information of the electronic mail is written in the person table in association with the information. 21. The person information generation means writes the sender's e-mail address acquired from the read header information of the e-mail as information specifying the sender in the person table. search support device according to any one of 22. Message identification information for identifying each electronic mail, thread identification information for identifying the thread to which the electronic mail belongs, a mail table composed of records including at least information on a sender and a receiver, and thread identification information for each thread And an analysis result storage means for storing a keyword table indicating information that associates the keywords with each other, presenting an important person related to the desired information, and a search support method used in a search support apparatus that supports the search of the information Because
The email reading means reads the email from the email storage means storing the email,
The mail analysis means acquires message identification information of other related electronic mail from the information of the header of the electronic mail read by the mail reading means, and whether the acquired message identification information is registered in the mail table. Judgment
When the analysis result output means determines that the mail analysis means is not registered, it newly generates thread identification information of the read e-mail, and when it is determined that it is registered, The thread identification information is acquired from the mail table record specified by the message identification information of other related emails, the message identification information acquired from the header information of the read email, the sender and the receiver, and newly A keyword that is a result of morphological analysis of information obtained from the read e-mail, and a record in which the generated thread identification information or thread identification information acquired from the mail table is set is added to the mail table, and the thread Write in the keyword table in association with the identification information,
The receiving means receives the search conditions and
The search means acquires thread identification information of a keyword that matches the search word indicated by the search condition received by the reception means from the keyword table, and is specified by the thread identification information for each acquired thread identification information. Obtaining a user set as a sender or a receiver from the record in the mail table and the number of the user set as a sender or a receiver,
Output control means, the number acquired by the searching means includes a step of outputting more user than a predetermined condition for each thread identification information,
The mail storage means sets a user who is a speaker in a conference as a sender, a participant of the conference as a receiver, a conference name of the conference as a subject, and a utterance content of the user as a speaker as a text. In addition, the message identification information of other e-mails that set other utterance contents in the meeting is set as the message identification information of other related e-mails including information indicating that it was generated based on the meeting. Remember emails,
The information indicating that the analysis result output means is generated based on the conference in the record to be added to the mail table when the information indicating that the analysis result output unit is generated based on the conference is set in the read e-mail. Set further,
The search means specifies, for each acquired thread identification information, each record set in the mail table specified by the thread identification information, and information indicating that the specified record is generated based on a meeting Is set, a predetermined point is given to the user set as the sender in each record, and a point lower than the predetermined point is given to the user set as the receiver. If the information indicating that it has been generated based on the above is not set, a point weighted more than the predetermined point is given to the user set as the sender and the user set as the receiver in each record And
The output control means outputs, for each thread identification information, information indicating a user whose total number of points given by the search means is greater than a predetermined condition.
A search support method characterized by that.   Message identification information for identifying each electronic mail, thread identification information for identifying the thread to which the electronic mail belongs, a mail table composed of records including at least information on a sender and a receiver, and thread identification information for each thread And an analysis result storage means for storing a keyword table indicating information that associates the keywords with each other, presenting an important person related to the desired information, and a search support method used in a search support apparatus that supports the search of the information Because
  The email reading means reads the email from the email storage means storing the email,
  The mail analysis means acquires message identification information of other related electronic mail from the information of the header of the electronic mail read by the mail reading means, and whether the acquired message identification information is registered in the mail table. Judgment
  When the analysis result output means determines that the mail analysis means is not registered, it newly generates thread identification information of the read e-mail, and when it is determined that it is registered, The thread identification information is acquired from the mail table record specified by the message identification information of other related emails, the message identification information acquired from the header information of the read email, the sender and the receiver, and newly A keyword that is a result of morphological analysis of information obtained from the read e-mail, and a record in which the generated thread identification information or thread identification information acquired from the mail table is set is added to the mail table, and the thread Write in the keyword table in association with the identification information,
  The receiving means receives the search conditions and
  The search means acquires thread identification information of a keyword that matches the search word indicated by the search condition received by the reception means from the keyword table, and is specified by the thread identification information for each acquired thread identification information. Obtaining a user set as a sender or a receiver from the record in the mail table and the number of the user set as a sender or a receiver,
  The output control means includes a step of outputting a user whose number acquired by the search means is greater than a predetermined condition for each thread identification information,

  The mail table record further includes broadcaster information;
  The analysis result output means adds, to the mail table, a record further setting information of the broadcaster acquired from the header information of the read electronic mail,
  For each acquired thread identification information, the search means performs predetermined management for a user set as a sender and a user set as a receiver in each record of the mail table specified by the thread identification information. And assigning an administrator search support point higher than the predetermined administrator search support point to the user set as a broadcaster,
  The output control means outputs, for each thread identification information, information indicating, as an administrator, a user whose total number of administrator search corresponding points is greater than a predetermined condition.
A search support method characterized by that. A computer used as a search support device that presents an important person related to desired information and supports search of the information,
Message identification information for identifying each electronic mail, thread identification information for identifying the thread to which the electronic mail belongs, a mail table composed of records including at least information on a sender and a receiver, and thread identification information for each thread And an analysis result storage means for storing a keyword table indicating information in which keywords are associated with each other,
Mail reading means for reading e-mail from mail storing means for storing e-mail;
Mail analysis means for acquiring message identification information of other related electronic mail from the information of the header of the electronic mail read by the mail reading means, and determining whether the acquired message identification information is registered in the mail table ,
If it is determined by the mail analysis means that it is not registered, thread identification information of the read e-mail is newly generated, and if it is determined that it is registered, the related other e-mail The thread identification information is acquired from the mail table record specified by the message identification information of the message, the message identification information acquired from the header information of the read electronic mail, the sender and the receiver, and the newly generated thread identification information or A record in which thread identification information acquired from the mail table is set is added to the mail table, and a keyword obtained as a result of morphological analysis of information obtained from the read electronic mail is associated with the thread identification information. Analysis result output means for writing to the keyword table;
Accepting means for receiving search conditions,
The mail table specified by the thread identification information is acquired for each thread identification information obtained from the keyword table, by acquiring thread identification information of a keyword that matches the search word indicated by the search condition received by the receiving means. Search means for obtaining a user who is set as a sender or a receiver from the record and a number of the user set as a sender or a receiver,
Output control means for outputting a user whose number acquired by the search means is greater than a predetermined condition for each thread identification information;
To function as,
The mail storage means sets a user who is a speaker in a conference as a sender, a participant of the conference as a receiver, a conference name of the conference as a subject, and a utterance content of the user as a speaker as a text. In addition, the message identification information of other e-mails that set other utterance contents in the meeting is set as the message identification information of other related e-mails including information indicating that it was generated based on the meeting. Remembers emails,
The information indicating that the analysis result output means is generated based on the conference in the record to be added to the mail table when the information indicating that the analysis result output unit is generated based on the conference is set in the read e-mail. Set further,
The search means specifies, for each acquired thread identification information, each record set in the mail table specified by the thread identification information, and information indicating that the specified record is generated based on a meeting Is set, a predetermined point is given to the user set as the sender in each record, and a point lower than the predetermined point is given to the user set as the receiver. If the information indicating that it has been generated based on the above is not set, a point weighted more than the predetermined point is given to the user set as the sender and the user set as the receiver in each record And
The output control means outputs, for each thread identification information, information indicating a user whose total number of points given by the search means is greater than a predetermined condition.
A computer program characterized by the above.   A computer used as a search support device that presents an important person related to desired information and supports search of the information,
  Message identification information for identifying each electronic mail, thread identification information for identifying the thread to which the electronic mail belongs, a mail table composed of records including at least information on a sender and a receiver, and thread identification information for each thread And an analysis result storage means for storing a keyword table indicating information in which keywords are associated with each other,
  Mail reading means for reading e-mail from mail storing means for storing e-mail;
  Mail analysis means for acquiring message identification information of other related electronic mail from the information of the header of the electronic mail read by the mail reading means, and determining whether the acquired message identification information is registered in the mail table ,
  If it is determined by the mail analysis means that it is not registered, thread identification information of the read e-mail is newly generated, and if it is determined that it is registered, the related other e-mail The thread identification information is acquired from the mail table record specified by the message identification information of the message, the message identification information acquired from the header information of the read electronic mail, the sender and the receiver, and the newly generated thread identification information or A record in which thread identification information acquired from the mail table is set is added to the mail table, and a keyword obtained as a result of morphological analysis of information obtained from the read electronic mail is associated with the thread identification information. Analysis result output means for writing to the keyword table;
  Accepting means for receiving search conditions,
  The mail table specified by the thread identification information is acquired for each thread identification information obtained from the keyword table, by acquiring thread identification information of a keyword that matches the search word indicated by the search condition received by the receiving means. Search means for acquiring a user set as a sender or a receiver from the record and a number of the user set as a sender or a receiver,
  Output control means for outputting a user whose number acquired by the search means is greater than a predetermined condition for each thread identification information;
  Function as
  The mail table record further includes broadcaster information;
  The analysis result output means adds, to the mail table, a record further setting information of the broadcaster acquired from the header information of the read electronic mail,
  For each acquired thread identification information, the search means performs predetermined management for a user set as a sender and a user set as a receiver in each record of the mail table specified by the thread identification information. And assigning an administrator search support point higher than the predetermined administrator search support point to the user set as a broadcaster,
  The output control means outputs, for each thread identification information, information indicating, as an administrator, a user whose total number of administrator search corresponding points is greater than a predetermined condition.
  A computer program characterized by the above.

Images