JP2012079158A - Retrieval device, retrieval program, and retrieval method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To highly rank document data really desired by a retriever in a retrieval result.SOLUTION: A mail data acquisition unit 11 acquires a sent e-mail with a communication history of the e-mail from a mail server 4. An expert degree calculation unit 12 specifies a transfer path of a sent e-mail of each sender from the communication history acquired by the mail data acquisition unit 11, and calculates an expert degree indicating expertise of the each sender for each word used by the each sender in the e-mail based on the specified transfer path. A score value calculation unit 13, when a creator of document data corresponds to the sender, increases the frequency of each word contained in each piece of the document data of the creator according to the corresponding expert degree of the word of the creator, thereby calculating a score value.

Description

  The present invention relates to a technique for retrieving document data.

  In recent years, in an in-house LAN or the like, document data created by each user is uploaded to the document server and the document data is released on the in-house LAN, and each user can freely access the document data stored in the document server from his terminal device. A system that can be browsed is introduced. Further, if the number of published document data becomes enormous, it will be difficult for each user to find desired document data. Therefore, in recent years, a search for causing each user to input a word related to the desired document data to the terminal device, searching the document data using the word from the document server, and transmitting the search result to the user terminal device. Equipment has also been introduced.

  In this case, the search device analyzes each word included in each document data stored in the document server, and calculates a score value in each document data for each word in advance. Then, when a word is input from the searcher, the search device displays, on the user terminal device, an image in which the file name of the document data in which the word is used is displayed in a list in descending order of score values for the word Let When the user clicks on a column of document data desired to be browsed, the document data is downloaded to the user's terminal device, and the user can browse the document data. As such a search device, for example, Patent Document 1 discloses a device that generates search candidates based on a search history of a server.

JP 2010-198577 A

  Hereinafter, problems with the conventional search device will be described. 18 to 20 are diagrams for explaining the problems of the conventional search device. For example, consider a case in which a user wants to view document data related to “CSR awareness activities”.

  In this case, the user inputs “csr enlightenment activity” as a word to be searched for in the word input field R1 at the upper right of FIG. 18, and clicks the search start button B1 on the right side of the word input field R1. Then, as shown in FIG. 18, the search result image in which the file name of the document data, the link destination of the document data, and the creator of the document data are listed in the descending order of the score value is the user terminal device. Is displayed.

  In FIG. 18, when the user clicks on the topmost document data column that is the first-order document data, the first-order document data is downloaded to the user's terminal device. Here, when browsing the first-ranked document data, the words “csr” and “enlightenment activity” were certainly included, but not the document data showing the detailed contents of “enlightenment activity” of “csr” It was not the document data that the searcher wanted.

  On the other hand, when the search result screen shown in FIG. 18 is scrolled downward, it can be seen that the document data related to “csr” is ranked in the 42nd position shown in FIG. Then, when this document data column was clicked and this document data was viewed, detailed contents about CSR enlightenment activities were described, and the document data desired by the searcher was found.

  FIG. 20 is an analysis of the search result of FIG. As shown in FIG. 20, the creator of the document data ranked in the 42nd place is (ΔΔΔΔ), and this creator is found to belong to the “CSR group”. Therefore, if the score value can be calculated in consideration of not only the word included in the document data but also the creator of the document data and the department to which the creator belongs, the document data ranked 42nd Document data desired by the searcher can be ranked higher by ranking higher.

  An object of the present invention is to provide a search device and the like that can rank document data that a searcher really wants at the top of search results by considering elements other than words included in the document data.

  (1) A search device according to the present invention is provided for each of a mail server for storing transmitted electronic mail, a terminal device used by a user, and a document server for storing document data to be disclosed on a predetermined network. A search device connected via the network, wherein the mail server stores the transmitted e-mail together with a communication history of each e-mail, and the search device is transmitted from the mail server. A mail data acquisition unit that acquires an e-mail together with the communication history, and a transfer path of the transmitted e-mail of each sender is specified from the communication history, and each sender is electronically based on the specified transfer path. For each word used in the email, each word included in each document data stored in the document server, and a degree-of-specialty calculating unit that calculates the degree of expertise indicating the specialty of each sender Based on the frequency, a score value calculation unit that calculates a score value for each document data of each word, and a search word input by the user using the terminal device is acquired from the terminal device, and is the same as the acquired search word Notification information for extracting the document data including the word from the document data stored in the document server, and notifying the extracted document data to the user in descending order of the score value for the same word The score value calculation unit is included in each document data of the creator when the creator of the document data corresponds to the sender. The score value is calculated by increasing the frequency of each word according to the degree of specialization of the corresponding word of the creator.

  In addition, the search program according to the present invention is provided for each of a mail server that stores transmitted electronic mail, a terminal device used by a user, and a document server that stores document data that is disclosed on a predetermined network. A search program for causing a computer to function as a search device connected via the network, wherein the mail server stores a transmitted electronic mail together with a communication history of each electronic mail, and the search program A mail data acquisition unit that acquires an electronic mail transmitted from a mail server together with the communication history, and a transmission path of the transmitted electronic mail of each sender is specified from the communication history, and based on the specified transfer path , For each word used by each sender in e-mail, Based on the frequency of each word included in each document data stored in the document server, a score value calculating unit that calculates a score value for each document data of each word, input by the user using the terminal device A search word is acquired from the terminal device, document data including the same word as the acquired search word is searched from the document server, and the searched document data is notified to the user in descending order of the score value. For generating the notification information for the user and transmitting the information to the terminal device of the user, the score function calculating unit, when the creator of each document data corresponds to the sender, The score value is calculated by weighting the frequency of each word included in each document data by the degree of specialization of the corresponding word of the creator.

  The search method according to the present invention provides a network for each of a mail server for storing transmitted electronic mail, a terminal device used by a user, and a document server for storing document data to be disclosed on a predetermined network. A search method using a search device connected via the mail server, wherein the mail server stores a transmitted electronic mail together with a communication history of each electronic mail, and the search device transmits from the mail server. A mail data acquisition step of acquiring a completed e-mail together with the communication history, and the search device specifies a transfer path of the transmitted e-mail of each sender from the communication history, and based on the specified transfer path And a degree-of-specialty calculation step of calculating a degree of specialization indicating the specialty of each sender for each word used by each sender in the e-mail, and the search device, A score value calculating step of calculating a score value for each document data of each word based on the frequency of each word included in each document data stored in the document server; The search word input using the terminal device is acquired from the terminal device, the document data including the same word as the acquired search word is searched from the document server, and the searched document data is searched in descending order of the score value. A search step for generating notification information for notification to the user and transmitting the notification information to the terminal device of the user, wherein the score value calculation step is performed when the creator of each document data corresponds to the sender, The score value is calculated by weighting the frequency of each word included in each document data of the creator by the degree of expertise of the corresponding word of the creator.

  According to this configuration, the degree of specialization of each word used by the sender in the email is calculated based on the transfer path of the email sent by the sender. In general, an e-mail transmitted by a key person who is a highly specialized user for a certain matter as a sender includes many keywords regarding the matter. In addition, the email created by the key person is highly important and tends to be forwarded by many users. Therefore, it can be estimated whether the sender of the e-mail is a key person from the scale of the transfer path. Therefore, in the present invention, the degree of specialization for each word used in the e-mail by the e-mail sender is calculated according to the e-mail transfer path.

  On the other hand, document data created by various creators is stored in the document server, and a score value is basically defined based on the frequency of each word included in each document data. Therefore, in the present invention, when the creator of the document data corresponds to the sender of the e-mail, the frequency of each word included in the document data is increased according to the specialty level of the corresponding word of the sender. Yes.

  Therefore, when a search word is input by the user, there is a high possibility that the score value of the document data created by the highly specialized user for the search word is high, not just the document data with a high frequency of the search word. .

  This makes it possible to rank document data created by a highly specialized user for the search word at the top of the notification information. Therefore, the user who has input the search word can easily access the document data that the user himself / herself wants to browse, in which the detailed contents of the search word are described.

  (2) The degree-of-specialization calculating unit has a first hierarchy of users who have directly received the e-mail from a sender of a certain e-mail, and a user who has received the e-mail transferred from the user of the first hierarchy It is preferable that the hierarchy consisting of is represented as the second hierarchy so that the transfer route of each electronic mail is represented hierarchically, and the degree of specialization is calculated higher as the hierarchy becomes deeper.

  In general, a highly important e-mail created by a key person is first sent from the key person to representatives of multiple departments in the organization having a pyramid-like organizational structure such as a company. There is a tendency to transmit in multiple hierarchies, such as being transmitted from representatives of each department to representatives of a plurality of sections belonging to each department, and being transmitted from representatives of each section to related members of each section. Therefore, as the number of hierarchies increases, the possibility that the e-mail is an e-mail created by a key person increases. Therefore, in the present invention, the higher the hierarchy, the higher the degree of specialization is calculated. As a result, it is possible to accurately estimate how much expertise the e-mail creator has for the words used in the e-mail, and reflect this in the degree of specialization.

  (3) It is preferable that the professional degree calculation unit calculates the special degree as the number of times of transfer of the e-mail increases.

  In general, highly important e-mails created by key persons tend to be forwarded to a large number of users by each forwarder. Therefore, in the present invention, by calculating the degree of specialization as the number of transfers increases, the e-mail creator accurately estimates the specialization about the word used in the e-mail and reflects it in the degree of specialization Can do.

  (4) The degree-of-profession calculation unit calculates, for a certain electronic mail, a product of a transfer depth indicating the depth of the hierarchy and the number of times the electronic mail is transferred from the hierarchy one level higher in each hierarchy. It is preferable that the degree of specialization for each mail level is calculated, and the sum or product of the calculated degrees of specialization for each level is calculated as the degree of specialization for each word included in the electronic mail.

  According to this configuration, as the product of the transfer depth and the transfer count increases, the degree of specialization is calculated. Therefore, the degree of specialization can be calculated such that the degree of specialization increases as the size of the transfer path increases.

  (5) When the certain sender transmits a plurality of e-mails and the same word is used in the plurality of e-mails, the professional degree calculation unit calculates the same word in each e-mail. It is preferable to calculate the total value of the expertise as the sender's expertise for the same word.

  According to this configuration, when a sender uses a word in a plurality of e-mails, the degree of specialization of the word can be accurately calculated.

  (6) It is preferable that the search unit generates, as the notification information, data for displaying a list of file names of searched document data in descending order of the score values.

  According to this configuration, since the image in which the file names of the document data are displayed as a list in the order of the highest score value is displayed on the terminal device, the user can recognize the document data with a high score value at a glance.

  According to the present invention, the degree of expertise for each word of the sender is calculated from the transfer path of the email, and the score value of each word of the document data is increased as the degree of specialty is increased. Document data can be ranked higher in the search results.

1 is an overall configuration diagram of a search system to which a search device according to an embodiment of the present invention is applied. It is a block diagram of the search device shown in FIG. It is the figure which showed an example of the data structure of a mail index. It is explanatory drawing of the calculation process of specialization. FIG. 5 is an explanatory diagram of a specialty degree calculation process for the first layer in the transfer path illustrated in FIG. FIG. 5 is an explanatory diagram of a degree of specialty calculation process for the second layer in the transfer path illustrated in FIG. FIG. 5 is an explanatory diagram of a degree of specialty calculation process for the third hierarchy in the transfer path illustrated in FIG. It is a figure which shows an example of the data structure of the specialty index which a specialty degree calculation part produces | generates. It is a figure which shows an example of the data structure of the score index which a score value calculation part calculates. It is the figure which showed an example of the data structure of the frequency index produced | generated by a document server. It is a block diagram of the terminal device shown in FIG. It is a block diagram of the document server shown in FIG. It is a block diagram of the mail server shown in FIG. It is the flowchart which showed the process of the search system when a search device calculates a score value. It is the flowchart which showed the process of the search system when a search device calculates a score value. It is the flowchart which showed the process of the search system at the time of the search of document data. It is the figure which summarized the process of the search system shown in FIG. It is a figure explaining the problem of the conventional search device. It is a figure explaining the problem of the conventional search device. It is a figure explaining the problem of the conventional search device.

  FIG. 1 is an overall configuration diagram of a search system to which a search device according to an embodiment of the present invention is applied. The search system shown in FIG. 1 includes a search device 1, one or a plurality of terminal devices 2, a document server 3, and a mail server 4. The search device 1 to the mail server 4 are connected via a predetermined network NT so that they can communicate with each other.

  In the present embodiment, an in-house communication network of a certain company can be adopted as the network NT, and for example, various data are transmitted using a TCP / IP communication protocol. As an in-house communication network, for example, the LAN of the head office, the LAN of each branch, the LAN of each sales office, the LAN of each laboratory, etc. A communication network connected to each other via a dedicated line such as 25 or VPN is adopted. Therefore, it is assumed that the network NT is not a global network such as the Internet but a private network.

  The terminal device 2 is composed of a general computer having a communication function, is possessed by a user who is a member of a company to which the present search system is applied, and creates various document data according to operations from each user. The created document data is uploaded to the document server 3 as necessary under the instruction of the user. Further, when a search command for document data related to a certain search word is input by the user, the terminal device 2 transmits a search request for causing the search device 1 to search for document data related to the word. To do. Further, the terminal device 2 creates an e-mail addressed to another user according to the user's operation, and transmits the created e-mail to the other user's terminal device 2 via the mail server 4.

  The search device 1 is composed of a general computer having a communication function. When a search request is received from the terminal device 2, the document data including the search word included in the search request among the document data stored in the document server 3. And the search result is displayed on the user terminal device.

  The document server 3 is composed of a general computer having a communication function, stores the document data uploaded from each terminal device 2, and receives a document data browsing request from a certain terminal device 2, the document data is stored in the document server 3. Download to the terminal device 2.

  The mail server 4 is composed of a general computer having a communication function. When an e-mail is transmitted from a certain terminal device 2, the e-mail server 4 stores the e-mail and stores it in the terminal device 2 of the user to whom the e-mail is transmitted. Send the email.

  FIG. 2 is a block diagram of the search device 1 shown in FIG. The search device 1 includes a mail data acquisition unit 11, a specialty level calculation unit 12, a score value calculation unit 13, a search unit 14, a specialty level storage unit 15, a score value storage unit 16, and a communication unit 17. Each of these blocks is realized by the CPU executing a search program for causing the computer to function as a search device.

  The search program is recorded on a computer-readable recording medium such as a DVD-ROM and distributed to the market. The search program is provided to the administrator by having the administrator of the search device 1 purchase the recording medium. Alternatively, the search program may be stored on an Internet Web server and downloaded to the administrator of the search apparatus 1 by downloading the search program.

  In FIG. 2, the mail data acquisition unit 11 to the search unit 14 are mainly configured by a CPU, and the specialty level storage units 15 and 16 are mainly configured by a large-capacity storage device such as an HDD (Hard Disk Drive). The communication unit 17 includes a communication module such as a modem or a LAN adapter.

  The mail data acquisition unit 11 acquires an electronic mail transmitted from the mail server 4 together with a communication history of the electronic mail. Here, the communication history is data indicating to which user the sender of the e-mail has transmitted the e-mail, and to which user the user who has received the e-mail has transferred the e-mail.

  In the present embodiment, the mail data acquisition unit 11 acquires the mail index shown in FIG. 3 from the mail server 4 via the communication unit 17. FIG. 3 is a diagram showing an example of the data structure of the mail index. The mail index shown in FIG. 3 is created for each electronic mail, and includes fields for a mail name, a sender, an electronic mail, and a communication history.

  In the mail name column, identification information for identifying the mail is stored, and in the present embodiment, for example, data including a set of a mail title and a transmission time is stored. For example, if the mail title is “about csr” and the transmission time is 13: 140: 30 on September 1, 2010, “about csr: 2010: 09: 01: 13: 40: 30” is identified. Adopted as information.

  In the sender column, identification information of the sender of the mail is described, and in this embodiment, the name and employee number of the sender of the mail are stored. The e-mail column stores the e-mail title and body data. If an attached file exists in the e-mail, the attached file is also stored.

  The communication history column stores an electronic mail communication history, and includes a transmission time, a sender, a receiver, and a reception time column. In the field of transmission time, the time when the sender sent the email, the time when the user who received the email from the sender became the sender (forwarder) and forwarded the email, and the forwarded email The time when the user who received the e-mail is further a sender (forwarder) and the e-mail is transferred is stored in time series.

  The sender field stores identification information of the sender of the email, the forwarder who forwarded the email from the sender, and the forwarder who further forwarded the forwarded email.

  The receiver column stores identification information of a receiver who directly receives an e-mail from a sender as a transmission source and a receiver who receives an e-mail from a forwarder.

  The reception time column stores the time at which each recipient stored in the receiver column receives the e-mail.

  In the example of FIG. 3, since N1 is stored in the first line of the sender column and N_a, N_b, N_c is stored in the first line of the receiver column, the user N1 becomes the sender and the user It can be seen that the electronic mail is transmitted to N_a, N_b, and N_c.

  In addition, since N_a is stored in the second row of the sender column and N_d, N_e is stored in the second row of the receiver column, the user N_a who receives the email from the user N1 is the user N_d, N_e. You can see that the email is being forwarded.

  Returning to FIG. 2, the professional degree calculation unit 12 identifies the transmission path of the transmitted email of each sender from the communication history of the mail index acquired by the mail data acquisition unit 11, and based on the identified transfer path For each word used by each sender in the e-mail, a degree of expertise indicating the specialty of each sender is calculated.

  Here, the degree-of-specialty calculation unit 12 includes a user who directly receives an e-mail from a sender of a certain e-mail as a first hierarchy, and a user who receives an e-mail transferred from a user in the first hierarchy. The hierarchy is referred to as the second hierarchy, and the transfer route of each e-mail is represented hierarchically, and the degree of specialization is calculated higher as the hierarchy becomes deeper. Further, the specialty level calculation unit 12 calculates the level of specialization as the number of times of transfer of electronic mail increases.

  FIGS. 4A and 4B are explanatory diagrams of the degree of specialty calculation process. FIG. 4A is a graph showing the transfer path of a certain e-mail e1, and FIG. 4B is a graph showing the degree of specialty. In the example of FIG. 4A, the sender N1 of the e-mail e1 is a user who belongs to the planning room of the head office. The sender N1 includes a user N_a who is a representative user of the branch manager room of the A branch, a user N_b who is a representative user of the branch manager room of the B branch, and a user N_c who is a representative user of the branch manager room of the C branch. The e-mail e1 is directly transmitted to the user with the name. The user N_a to whom the e-mail e1 is transmitted determines that the e-mail e1 needs to be transferred, and the e-mail e1 is sent to the user N_d who is the representative user of the general affairs section of the A sales office and the It is transferred to the user N_e who is the representative user.

  The user N_d to whom the e-mail e1 has been transferred from the user N_d determines that the e-mail e1 needs to be transferred, and transfers the e-mail e1 to the users N_A, N_B, N_C who are employees of the A salesperson. In this way, when the transfer route of the e-mail e1 is traced, the transfer route of the e-mail e1 is represented hierarchically. Then, in the transfer path shown in FIG. 4A, the professional degree calculation unit 12 sets the hierarchy to which the sender N1 belongs to the reference hierarchy, and the hierarchy to which the users N_a, N_b, and N_c that have received the e-mail e1 directly from the sender belong. The first hierarchy, the hierarchy to which the user who has received the e-mail e1 transferred from the user in the first hierarchy belongs to the second hierarchy, and the hierarchy to which the user who has received the e-mail e1 transferred from the user in the second hierarchy belongs to the third hierarchy. The hierarchy is defined such that the hierarchy to which the user who has received the e-mail e1 transferred from the user of the hierarchy and the (i-1) th hierarchy belongs is called the i-th hierarchy.

  That is, the specialty level calculation unit 12 increases the hierarchy as the distance from the sender N1 increases in the transfer route illustrated in FIG. Then, the specialty level calculation unit 12 defines the transfer depth by setting the transfer depth of the first hierarchy to 1, the hierarchy depth of the second hierarchy to 2, and the transfer of each hierarchy as shown in FIG. The product of the depth and the number of transfers of the e-mail e1 at each level is calculated as the degree of specialization for each level of the e-mail e1.

  FIG. 5 is an explanatory diagram of the process of calculating the degree of specialization for the first layer in the transfer route shown in FIG. 4A. FIG. 5A shows the same transfer route as FIG. FIG. 5 is a graph showing a process of calculating the degree of specialization in the first hierarchy in FIG.

  As shown in FIG. 5A, since the sender N1 has transmitted the e-mail e1 to the three users N_a to N_c, the number of transfers from the reference layer to the first layer is 3. Further, the users N_a to N_c belong to the first layer, and the transfer depth of the first layer is 1. Therefore, the professional level calculation unit 12 calculates the professional level DV_1 in the first hierarchy as DV_1 = 1 × 3 = 3 as shown in FIG. 5B.

  FIG. 6 is an explanatory diagram of the process of calculating the degree of specialization for the second layer in the transfer route shown in FIG. 4A. FIG. 6A shows the same transfer route as FIG. FIG. 4 is a graph showing a process of calculating the specialty level DV_2 of the second hierarchy in FIG.

  In the example of FIG. 6A, it is assumed that the users N_a, N_b, and N_c have transferred the e-mail e1 to two users such as the users N_d and N_e, respectively. Therefore, the number of transfers from the first layer to the second layer is 2 × 3 = 6. The six users to whom the e-mail e1 is transferred from the users N_a, N_b, and N_c belong to the second hierarchy, and the transfer depth of the second hierarchy is 2. Therefore, as shown in FIG. 6B, the professional level calculation unit 12 calculates the professional level DV_2 as DV_2 = 2 × 6 = 12.

  FIG. 7 is an explanatory diagram of the process of calculating the total specialty level of the first to third layers in the transfer route shown in FIG. 4A. FIG. 7A shows the same transfer route as that in FIG. (B) is a graph showing the calculation process of the total specialty DV of the first to third hierarchies in FIG. 4 (A).

  In the example of FIG. 7A, it is assumed that the users N_d and N_e have respectively transferred the e-mail e1 to three users. Therefore, the number of transfers from the second layer to the third layer is 3 × 2 = 6. Further, the user to whom the e-mail e1 is transferred from the users N_d and N_e belongs to the third hierarchy, and the transfer depth of the third hierarchy is 3. Therefore, the professional level calculation unit 12 calculates the professional level DV_3 of the third hierarchy as DV_3 = 6 × 3 = 18.

  Then, the specialty level calculation unit 12 calculates 33 (= 3 + 12 + 18), which is the total value of the specialty levels DV_1, DV_2, and DV_3, as the expertise level DV of the sender N as illustrated in FIG.

  In this way, the professional level calculation unit 12 calculates the product of the number of transfers from the (i-1) th layer to the i-th layer and i, which is the transfer depth of the i-th layer, as the professional level DV_i of the i-th layer, The total value of the level of expertise DV_i is calculated as the level of expertise of the sender N.

  That is, the professional level calculation unit 12 calculates the professional level DV using the formula (1).

DV = Σ _{i = 1} ^last DV_i (1)
DV_i = i · F (i)
However, F (i) represents the number of transfers from the (i-1) th layer to the ith layer, and last represents the maximum value of the transfer depth.

  In addition, although the professional level calculation part 12 calculated total professional level from the sum of the professional level of each hierarchy, it is not limited to this, You may calculate total professional level from the product of the professional level of each hierarchy . By doing so, the difference in expertise between senders can be made more apparent.

  Then, the specialty level calculation unit 12 extracts words included in the e-mail e1 from the data stored in the e-mail column of the e-mail index of the e-mail e1 acquired by the mail data acquisition unit 11, and extracts each extracted word Is assigned with the degree of specialty DV calculated for the e-mail e1. For example, if the e-mail e1 includes the words W1 to W3, and the e-mail e1 has a special degree of DV_e1, the specialities of the words W1 to W3 are calculated to be DV_e1, respectively.

  Here, when a certain sender transmits a plurality of e-mails and the same word is used in the plurality of e-mails, the specialty level calculation unit 12 calculates the expertise calculated in each e-mail for the same word. The total value of degrees may be calculated as the degree of expertise of the sender for the same word.

  For example, it is assumed that the sender N has sent e-mails e1 and e2. Assume that the e-mails e1 and e2 both include the word W1. Assume that the specialty level DV_e1 of the e-mail e1 is calculated as α_e1, and the specialty level DV_e2 of the e-mail e2 is calculated as α_e2. In this case, the professional degree calculation unit 12 calculates the professional degree α1 regarding the word W1 of the sender N as α1 = α_e1 + α_e2.

  FIG. 8 is a diagram illustrating an example of a data structure of the professional degree index generated by the professional degree calculation unit 12. As shown in FIG. 8, the specialty index is generated individually for each user. In the example of FIG. 8, the specialty index of the users N1 and N2 is shown.

  The specialty index includes columns for words and expertise. In the word column, words used in e-mails transmitted by the user in the past are described. In the specialty level column, the specialty level of each word described in the word column is described.

  For example, since the user N1 uses the words W1 to W3 in the electronic mail, W1 to W3 are described in the word column. Further, since the specialities of the words N1 to W3 of the user N1 are calculated as α1, α2, and α3, respectively, the specialities of α1, α2, and α3 are described in the speciality column for the words W1 to W3. Yes.

  In addition, since the user N2 uses the words W3 to W5 in the electronic mail, W3 to W5 are described in the word column of the specialty index of the user N2. Further, since the specialities of the words W3 to W5 of the user N2 are calculated as β1, β2, and β3, respectively, the specialities of β1, β2, and β3 are described in the speciality column for the words W3 to W5. Yes.

  In this way, the professional level calculation unit 12 calculates the professional level for each email sent by each user in the past, classifies the calculated professional level for each word, and calculates the total value of the professional level for each word. By doing so, the specialty index of each user shown in FIG. 8 is calculated.

  In the above description, the degree of specialization DV_e1 is uniformly assigned to each word included in the e-mail e1 regardless of the frequency of each word. However, the present invention is not limited to this. The degree of specialization may be calculated individually for each word.

  In this case, for example, the professional level calculation unit 12 may obtain the frequency ver_e1 of each word included in the e-mail e1 and set a value obtained by weighting the professional level DV_e1 with the frequency ver_e1 as the professional level of each word.

  Returning to FIG. 2, the score value calculation unit 13 calculates a score value for each document data of each word based on the frequency of each word included in each document data stored in the document server 3. Here, when the creator of the document data corresponds to the sender, the score value calculation unit 13 determines the frequency of each word included in each document data of the creator according to the specialty level of the corresponding word of the creator. The score value is calculated by increasing the value.

  For example, document data D1 of the user N1 is stored in the document server 3, and the document data D1 includes words W1 and W4. The frequency of the word W1 is ver_W1 and the frequency of the word W4 is ver_W4. .

  On the other hand, in the professional degree index of the user N1 shown in FIG. 8, the professional degree of the word W1 is calculated as α1, and the professional degree of the word W4 is not calculated. In this case, the score value calculation unit 13 calculates the score value SR_D1_W1 of the word W1 in the document data D1 as SR_D1_W1 = ver_W1 × α1, and calculates the score value SR_D1_W4 of the word W4 in the document data D1 as SR_D1_W4 = ver_W4 × 1. To do. Here, since the specialization degree α1 is obtained by the product of the transfer depth and the transfer count, it has a value of 1 or more. Accordingly, the score value SR_D1_W1 has a larger value as the specialty degree α1 increases as a result of the frequency ver_W1 being increased according to the specialty degree α1.

  Then, the score value calculation unit 13 calculates the score value of each word for each document data stored in the document server 3, and generates the score index shown in FIG.

  FIG. 9 is a diagram illustrating an example of the data structure of the score index calculated by the score value calculation unit 13. The score index is created for each document data. In the example of FIG. 9, the score indexes of the document data D1 and D2 are shown. As shown in FIG. 9, the score index includes columns for document name, creator, word, and score value.

  In the document name column, document data identification information is stored, and in the present embodiment, the file name of the document data is stored. In the creator column, the identification information of the user who created the document data is stored, and in this embodiment, the user name is stored. In the word column, words included in the document data are stored. The score value column stores the score values calculated by the score value calculation unit 13 for each word stored in the word column.

  In the present embodiment, as shown in FIG. 10, the document server 3 creates a frequency index for each document data, and the score value calculation unit 13 obtains the frequency index from the document server 3 via the communication unit 17. The score index shown in FIG. 9 is created. By doing so, the amount of data transmitted and received between the document server 3 and the search device 1 is suppressed as compared with the case where the search device 1 receives the document data itself from the document server 3 and creates a score index. Can do.

  FIG. 10 is a diagram showing an example of the data structure of the frequency index generated by the document server 3. The frequency index shown in FIG. 10 is the same as the score index shown in FIG. 9 except for the frequency column. The frequency column of the frequency index stores the frequency of each word included in the document data. Here, as the frequency, for example, the number of times the word is used in the document data may be adopted, or the ratio of the number of times each word is used with respect to the word that is used the maximum number of times in the document data. Also good.

  Returning to FIG. 2, the search unit 14 acquires a search word input by the user X using the terminal device 2 from the terminal device 2 via the communication unit 17, and includes document data including the same word as the acquired search word Is extracted from the document data stored in the document server 3, and notification information for notifying the extracted document data to the user X in the descending order of score values for the same word is generated. It transmits to the terminal device 2.

  Here, as the notification information, data that displays a list of file names of the retrieved document data in descending order of score values is generated as the notification information.

  For example, it is assumed that the word W1 is input by the user X as a search word. In this case, the search unit 14 acquires the word W1 via the communication unit 17, refers to the score index shown in FIG. 9, extracts document data that includes the same word as the word W1, and extracts the extracted document data Among them, HTML data for displaying a list of file names of a predetermined number of document data having a high score value for the word W1 in descending order of the score value is generated and passed to the communication unit 17. Then, the communication unit 17 transmits the passed HTML data to the terminal device 2 of the user X.

  In this way, in FIG. 18, in the document data search list when “csr enlightenment activity” is entered as a search word, the document data of “Monthly CSR communication” displayed in the 42nd position is listed first. The possibility of displaying can be increased.

  The search unit 14 refers to the score index stored in the score value storage unit 16 to extract document data including the same word as the search word. Therefore, the search unit 14 directly accesses the document server 3 to search for the search word. The amount of data transmitted and received between the search device 1 and the document server 3 can be reduced compared to extracting document data including the same word.

  Returning to FIG. 2, the specialty degree storage unit 15 stores the specialty degree index (see FIG. 8) generated by the specialty degree calculation unit 12. The score value storage unit 16 stores the score index generated by the score value calculation unit 13 (see FIG. 9).

  The communication unit 17 determines whether a packet flowing through the network NT is a packet addressed to the search device 1. If the packet is addressed to the search device 1, the communication unit 17 extracts necessary data from the packet, and a block that requires the data. To pass. Further, when there is a data transmission request from the search unit 14 or the like, the communication unit 17 receives the data, generates a packet conforming to the TCP / IP communication protocol, and transmits the packet to the network NT.

  In the present embodiment, when receiving the mail index (see FIG. 3) transmitted from the mail server 4, the communication unit 17 passes the mail index to the mail data acquisition unit 11. Further, when receiving the frequency index (see FIG. 10) transmitted from the document server 3, the communication unit 17 passes the frequency index to the score value calculation unit 13. Further, when receiving a search request including a search word from the terminal device 2, the communication unit 17 passes the search request to the search unit 14. In addition, when the notification information is passed from the search unit 14, the communication unit 17 transmits the notification information to the corresponding terminal device 2.

  FIG. 11 is a block diagram of the terminal device 2 shown in FIG. The terminal device 2 includes an operation unit 21, a control unit 22, a display unit 23, and a communication unit 24. The operation unit 21 includes input devices such as a keyboard and a mouse, and accepts various operation commands from the user.

  The control unit 22 includes a CPU, a ROM, a RAM, a hard disk, and the like, and executes various processes according to operation commands from the user input using the operation unit 21. In the present embodiment, the control unit 22 generates document data in accordance with an operation command from the user, and when an instruction to upload the document data to the document server 3 is input, the control unit 22 sends the document data to the communication unit 24. hand over.

  Further, the control unit 22 generates an e-mail according to an operation command from the user, and when the e-mail transmission command is input, passes the e-mail to the communication unit 24.

  In addition, when a search word is input from a user and a search command for document data is input, the control unit 22 passes a search request for document data including the search word to the communication unit 24.

  The display unit 23 includes a display device such as a liquid crystal display, a CRT, and a plasma display.

  The communication unit 24 includes a communication module such as a modem or a LAN adapter. In the present embodiment, when document data is transferred from the control unit 22, the communication unit 24 transmits the document data to the document server 3. In addition, when the electronic mail is delivered from the control unit 22, the communication unit 24 transmits the electronic mail to the mail server 4. In addition, when a search request is passed from the control unit 22, the communication unit 24 transmits the search request to the search device 1.

  FIG. 12 is a block diagram of the document server 3 shown in FIG. The document server 3 includes a communication unit 31, a control unit 32, and a document data storage unit 33. The communication unit 31 includes a communication module such as a modem or a LAN adapter. When the document data uploaded from the terminal device 2 is received, the communication unit 31 passes the document data to the control unit 32. In addition, when the frequency index is passed from the control unit 32, the communication unit 31 transmits the frequency index to the search device 1. Further, when receiving a document data browsing request from a certain terminal device 2, the communication unit 31 passes the browsing request to the control unit 32.

  The control unit 32 includes a CPU, a ROM, a RAM, and the like, and stores the uploaded document data received by the communication unit 31 in the document data storage unit 33. Also, when a request for browsing certain document data is passed from a certain terminal device 2 by the communication unit 31, the control unit 32 reads the document data from the document data storage unit 33, passes the document data to the communication unit 31, and applies the document data to the corresponding unit. To the terminal device 2 to be transmitted.

  Further, the control unit 32 generates a frequency index (see FIG. 10) for each document data stored in the document data storage unit 33, passes the frequency index to the communication unit 31, and passes the frequency index to the search device 1. Send. Here, the control part 32 extracts a word about each document data, and produces | generates a frequency index by counting the frequency | count of use of the extracted word. When the document data is Japanese document data, the control unit 32 may extract mainly nouns, verbs, adjectives, adjective verbs, and the like as words, excluding particles and conjunctions. When the document data is document data in English, French, German, etc., the control unit 32 mainly extracts nouns, verbs, adjectives, adverbs, etc. as words, excluding definite articles, articles, prepositions, etc. That's fine.

  The document data storage unit 33 includes a storage device such as a hard disk, and stores document data uploaded by the user.

  FIG. 13 is a block diagram of the mail server 4 shown in FIG. The mail server 4 includes a communication unit 41, a control unit 42, and a mail data storage unit 43. The communication unit 41 is configured by a communication module such as a modem or a LAN adapter.

  The control unit 42 includes a CPU, a ROM, a RAM, and the like. When an e-mail transmitted from the terminal device 2 is received by the communication unit 41, the control unit 42 stores the e-mail in the mail data storage unit 43. The electronic mail reception notification is passed to the communication unit 41 to the terminal device 2 of the user of the electronic mail destination, and the reception notification is transmitted to the terminal device 2 of the destination user.

  When the communication unit 41 receives an email reception request from the terminal device 2 of the user that is the destination of the email, the control unit 42 passes the email to the communication unit 41 and passes the email to the destination. It transmits to the terminal device 2 of the user.

  In addition, the control unit 42 specifies the identification information of the email, the sender, the recipient, and the transmission time from the data included in the header of the email, and also sends the email from the terminal device 2 of the destination user. The reception time of the electronic mail is specified from the reception time of the reception request, and the mail index shown in FIG. 3 is generated from these specified data and stored in the mail data storage unit 43. In addition, the control unit 42 passes the generated mail index to the communication unit 41 as necessary, and transmits it to the search device 1.

  The mail data storage unit 43 includes a storage device such as a hard disk, and stores electronic mail and a mail index.

  14 and 15 are flowcharts showing the processing of the search system when the search device 1 calculates the score value. First, when the search device 1 reaches a predetermined time (YES in step S11), the search device 1 transmits a mail index transmission request to the mail server 4 (step S12). Here, when the score value is calculated every day, a certain time such as 0:00 is adopted as the predetermined time, and when the score value is calculated every week, a certain day of the week such as 0:00 every Sunday When a score value is calculated every month, a time on a certain day such as 9 o'clock on the 25th of every month is employed. Note that the calculation interval of the score value is not limited to one day, one week, and one month, and a suitable interval such as February and June may be adopted as appropriate.

  If the predetermined time has not been reached in step S11 (NO in step S11), the process returns to step S11.

  Upon receiving the mail index transmission request (step S1), the mail server 4 transmits the mail index stored in the mail data storage unit 43 and generated in advance to the search device 1 (step S2). Here, the mail server 4 may transmit only the unsent mail index to the search device 1, that is, the mail index difference data may be transmitted to the search device 1.

  When the search device 1 receives the mail index (step S13), the search device 1 identifies the transfer route of each email from the communication history stored in the email index of each email (step S14).

  In the example of the communication history in FIG. 3, the sender of the transmission source is the user N1, the user N1 transfers the email to the users N_a, N_b, and N_c, and the user N_a sends the email to the users N_d and N_e. Forwarding. Therefore, the search device 1 generates a transfer path in which the user N1 is the reference hierarchy, the users N_a, N_b, and N_c are the first hierarchy, and the users N_d and N_e are the second hierarchy.

  Next, the search device 1 calculates the specialty level of each e-mail from the transfer path specified for each e-mail (step S15). Next, the search device 1 extracts a word for each e-mail, and calculates the expertise calculated in step S15 as the expertise of each extracted word (step S16).

  Next, the search device 1 generates a specialty index (see FIG. 8) by classifying the expertise of each word calculated for each e-mail calculated in step S16 by sender (see FIG. 8). In this case, in the case where the same sender transmits a plurality of e-mails and the same word is included in the plurality of e-mails, the search device 1 adds up the expertise of these same words. By doing so, the degree of specialization for each word of each sender is calculated.

  Next, the search device 1 transmits a frequency index transmission request to the document server 3 (step S18). Upon receiving the frequency index transmission request (step S31), the document server 3 extracts the creator and the word from each document data stored in the document data storage unit 33, and obtains the frequency of the extracted word in each document data. Then, a frequency index (see FIG. 10) for each document data is generated (step S32). In this case, when there is document data for which a frequency index has already been generated, the document server 3 does not generate a frequency index for the document data, and only generates a frequency index for document data for which no frequency index has been generated.

  Next, the document server 3 transmits the frequency index generated in step S32 to the search device 1 (step S33).

  When the search device 1 receives the frequency index (step S19), the search device 1 calculates the score value of each word for each document data using the received frequency index and the specialty index generated in step S17, and the score index (See FIG. 9) is generated (step S20). In this case, the search device 1 may generate a score index only for document data for which a score index has not been generated. Further, even in the case of document data for which the score value has been calculated, when the level of expertise of the creator of the document data is changed, the level of expertise of each word in the document data may be calculated again. Thereby, a score index having a score value reflecting the latest degree of specialization can be calculated.

  Next, the search device 1 stores the score index generated in step S20 in the score value storage unit 16 (step S21).

  FIG. 16 is a flowchart showing the processing of the search system when searching for document data. First, when a search word is input by the user X and a search command for document data including the search word is input from the user X (YES in step S51), the terminal device 2_X of a certain user X sends a search request to the search device 1. (Step S52).

  On the other hand, if no search command is input in step S51 (NO in step S51), the process returns to step S51.

  When receiving the search request (step S61), the search device 1 refers to the score index stored in the score value storage unit 16 to extract document data including the search word included in the search request (step S62). .

  Next, the search device 1 further extracts a predetermined number of document data in descending order of score values for the search word from the extracted document data, and extracts the file name, creator, and link destination of the extracted predetermined number of document data. Etc. are generated in order of the score values for the search words in descending order (step S63).

  Next, the search device 1 transmits the notification information generated in step S63 to the terminal device 2_X (step S64). When the terminal device 2_X receives the notification information (step S53), the terminal device 2_X displays, on the display unit 23, an image in which the file name of the document data is displayed in a list according to the notification information (step S54).

  Thereby, it is possible to provide the user X with a search result in which document data created by a highly specialized user is ranked higher for the search word input by the user X.

  Next, when any document data column is clicked in the image displayed as a list from the user X and the terminal device 2_X receives a browsing instruction for the document data (YES in step S55), the terminal device 2_X sends the browsing instruction to the document server 3. (Step S56). On the other hand, when a browsing instruction is not input (NO in step S55), the process returns to step S55.

  When receiving the browsing instruction (step S41), the document server 3 transmits the document data that the user X desires to browse to the terminal device 2_X (step S42).

  When the terminal device 2_X receives the document data (step S57), the terminal device 2_X displays the document data on the display unit 23 (step S58). As a result, the user X can easily browse the document data describing the desired content for the search word.

  FIG. 17 is a summary of the processing of the search system shown in FIG. First, in step ST1, a sender N1 who is in charge of CSR transmits an e-mail e1 including a CSR word. Hereinafter, it is assumed that the specialty level of the sender N1 for the CSR word is calculated as 33. Further, the sender N1 creates document data D1 including the word of CSR and uploads it to the document server 3 (step ST2). Accordingly, the score value of the CSR word included in the document data D1 of the sender N1 is a value obtained by multiplying the frequency by 33 (step ST3).

  On the other hand, the user N_A creates document data D2 including the word of CSR, uploads it to the document server 3 (step ST4), and transfers the e-mail e1 (step ST5).

  However, since the user N_A is not the sender of the e-mail e1, the score value of the CSR word included in the document data D2 is calculated as frequency × 1 and stored in the score index (step ST6).

  Therefore, when CSR is input as a search word, the search result is obtained in which the document data D1 created by the person in charge of CSR is ranked higher than the document data D2 created by the user who is not the person in charge of CSR ( Step ST7).

  In this way, the degree of expertise of each word used by the sender in the email is calculated based on the transmission path of the sender's sent email, and the score value is calculated using the degree of expertise. Document data created by a user who has high expertise regarding the search word can be ranked at the top of the search results.

DESCRIPTION OF SYMBOLS 1 Search apparatus 2 Terminal apparatus 3 Document server 4 Mail server 11 Mail data acquisition part 12 Specialty degree calculation part 13 Score value calculation part 14 Search part 15 Specialty degree storage part 16 Score value storage part 17 Communication part DV Specialty degree NT network

Claims (8)

A search connected via the network to each of a mail server for storing transmitted electronic mail, a terminal device used by a user, and a document server for storing document data to be disclosed on a predetermined network A device,
The mail server stores the sent e-mail together with the communication history of each e-mail,
The search device includes:
A mail data acquisition unit for acquiring an email sent from the mail server together with the communication history;
A specialized transfer path for each sender's sent e-mail is identified from the communication history, and each sender's e-mail is based on the specified transfer path. A degree-of-specialty calculator that calculates the degree,
A score value calculation unit that calculates a score value for each document data of each word based on the frequency of each word included in each document data stored in the document server;
A search word input by the user using the terminal device is acquired from the terminal device, and document data including the same word as the acquired search word is extracted from the document data stored in the document server, and extracted. A search unit that generates notification information for notifying the user of the document data in descending order of the score value for the same word, and transmitting the information to the terminal device of the user,
When the creator of the document data corresponds to the sender, the score value calculation unit determines the frequency of each word included in each document data of the creator according to the degree of specialty of the corresponding word of the creator The score is calculated by increasing the score value.   The degree-of-specialization calculating unit includes a first layer including a user who directly receives an e-mail from a sender of a certain e-mail, and a layer including a user who receives the e-mail transferred from the user of the first layer. The search apparatus according to claim 1, wherein the e-mail transfer path is hierarchically expressed as a second hierarchy, and the degree of specialization is calculated higher as the hierarchy becomes deeper.   The search device according to claim 1, wherein the specialty level calculation unit calculates the specialty level higher as the number of times of transfer of the e-mail increases.   The degree-of-specialization calculating unit calculates a product of a transfer depth indicating a depth of the hierarchy and a transfer count of the email from a hierarchy one level higher in each hierarchy for a certain email. The search device according to claim 2, wherein a degree of specialization is calculated for each word, and a sum or product of the calculated degrees of specialization is calculated as the degree of specialization for each word included in the electronic mail.   The degree-of-specialization calculating unit transmits a plurality of e-mails, and when the same word is used in the plurality of e-mails, the degree-of-specialty calculated in each e-mail for the same word The search device according to any one of claims 1 to 4, wherein a total value is calculated as the degree of expertise of the sender for the same word.   The search device according to claim 1, wherein the search unit generates, as the notification information, data that displays a list of file names of searched document data in descending order of the score values. A search connected via the network to each of a mail server for storing transmitted electronic mail, a terminal device used by a user, and a document server for storing document data to be disclosed on a predetermined network A search program for causing a computer to function as a device,
The mail server stores the sent e-mail together with the communication history of each e-mail,
The search program is
A mail data acquisition unit for acquiring an email sent from the mail server together with the communication history;
A specialized transfer path for each sender's sent e-mail is identified from the communication history, and each sender's e-mail is based on the specified transfer path. A degree-of-specialty calculator that calculates the degree,
A score value calculation unit that calculates a score value for each document data of each word based on the frequency of each word included in each document data stored in the document server;
A search word input by the user using the terminal device is acquired from the terminal device, document data including the same word as the acquired search word is searched from the document server, and the searched document data is Generate notification information for notification to the user in descending order of score value, function the computer as a search unit to transmit to the terminal device of the user,
When the creator of each document data corresponds to the sender, the score value calculation unit weights the frequency of each word included in each document data of the creator by the specialty level of the corresponding word of the creator A search program for calculating the score value. A search connected via the network to each of a mail server for storing transmitted electronic mail, a terminal device used by a user, and a document server for storing document data to be disclosed on a predetermined network A search method using a device,
The mail server stores the sent e-mail together with the communication history of each e-mail,
A mail data acquisition step in which the search device acquires an email sent from the mail server together with the communication history;
The search device identifies the transmission path of each sender's transmitted e-mail from the communication history, and for each word used by the sender in the e-mail based on the identified transfer path, A degree-of-profession calculation step for calculating a degree of specialty indicating expertise,
A score value calculating step in which the search device calculates a score value for each document data of each word based on the frequency of each word included in each document data stored in the document server;
The search device acquires a search word input by the user using the terminal device from the terminal device, searches the document server for document data including the same word as the acquired search word, and searches A search step for generating notification information for notifying the user of document data in descending order of the score value, and transmitting the information to the terminal device of the user,
In the score value calculating step, when the creator of each document data corresponds to the sender, the frequency of each word included in each document data of the creator is weighted by the specialty level of the corresponding word of the creator The search method which calculates the said score value by doing.