JP2008077250A - Information processing method, information processor, program and recording medium which records this - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an information processing method which can automatically distinguish same events and similar events, and generate and update schedule information, an information processor, a program for information processing, and a recording medium which records this. <P>SOLUTION: This method is provided with; a stage for acquiring a message via a mail or a bulletin board on the Internet or an intranet by a message acquisition section; a stage for extracting information which expresses a specific event using a syntactic structure from a message which is acquired by an event information extraction section; a stage for determining the same event and the similar event from the event information extracted by a schedule creation section and automatically generate the schedule; and a stage for outputting the schedule which is created by a schedule output section. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to an information processing system that includes information extraction from text and information editing, and in particular, an information processing method, an information processing apparatus, and an information processing method capable of automatically determining the same event or similar event to generate and update schedule information. The present invention relates to a program for information processing and a recording medium on which the program is recorded.

In recent years, with the development of electronic communication such as e-mail and electronic bulletin boards, opportunities to adjust schedules of groups and individuals using e-mails and electronic bulletin boards have increased, and the spread of Personal Information Manager (PIM) software has become popular. Along with this, an increasing number of users manage the adjusted schedule with a computer.
This PIM software is known to have a schedule management function that displays a calendar format on a display device of a terminal such as a computer, for example, and allows the user to store and manage the schedule of the day on the date of the displayed calendar. .
However, conventionally, it is necessary for the user to manually input the adjusted schedule, and the cooperation between the electronic communication and the PIM software is not sufficient.

The technology for automatically extracting useful information from documents and editing and outputting it in a form that is easy to use has advanced. For the purpose of reflecting event information extracted from electronic communication means in personal schedule management software, the first In the prior art, there is proposed a schedule information extraction / editing apparatus capable of automatic extraction of schedule information, automatic reflection to a scheduler, and further editing using the relationship between messages. The schedule information is automatically extracted using direct pattern matching or pattern matching to the morpheme analysis result to generate a schedule registration candidate. Here, the schedule determination process that determines whether or not the schedule is necessary according to the extraction rule specified by the user, the addition of a schedule that could not be automatically extracted while referring to the message, and the relationship between messages provided by a certain guideline It is possible to perform an editing process such as deleting a schedule with reference to the strength or weakness of the document (Patent Document 1).
In the second prior art, a method for storing a mail document in association with the date of the scheduler, and a mechanism for specifying which of the actual event content date and time in the mail and the mail reception date to be associated with can be specified. It has been proposed (Patent Document 2).
JP 2004-102559 A JP 2005-322178 A

However, regarding the automatic extraction of schedule information in the first prior art, there is a problem in that information to be registered in the schedule cannot be specified if the target message includes a plurality of notations that match the date and time pattern. is there.
In addition, it is cumbersome and costly for the user to set the extraction rules required for the schedule determination process, and manually edit the schedule, such as adding and deleting schedules while referring to the message body and message relationships. Absent.

Further, in the second prior art, the message can be associated with the scheduler, but the association of the message attached document is not mentioned. In addition, it is necessary for the user to intervene in specifying the date for association, and it takes time and cost to register the schedule.
Some of the acquired messages are related to the same event information, and there are also events in which information is added as time passes. For example, a conference event is considered to have a schedule adjustment message and a holding guidance message. If a schedule is automatically generated for each message, the amount of the schedule becomes enormous and management reference becomes difficult. Therefore, it is desirable that the same event can be handled as the same schedule.

  The message document may be expressed differently even in the content of the same event. For example, between events in which event elements are almost the same, an event element expression “When in the morning” and “10:00” Although the expression of the event element “from 12 o'clock” indicates the same time zone, it cannot be determined conventionally. If it can be determined that they are in the same time zone, these two events can be generated as one schedule as the same event.

It is also possible that the minutes are written after the meeting is over, and that they are sent and acquired as messages. Also at this time, it is preferable that the schedule is handled as the same schedule and the schedule is automatically updated as time passes, rather than being automatically generated as another event. In the conventional method, there is no mention of an effective method for dealing with this problem.
Here, in these prior arts, there is no mention of document attachment to the schedule, but an image document of a guide map attached to an exhibition notification e-mail or the like is attached to the schedule or a conference draft There is a high demand for users to attach a schedule or minutes as a related document to the schedule and refer to it as necessary.

  The present invention is for solving the above-mentioned conventional problems, and its purpose is to automatically extract event information from a message transmitted daily by electronic communication means using an information extraction technique using a syntactic structure. By providing an information processing method, an information processing apparatus, a program for information processing, and a recording medium on which this information is recorded, which can automatically determine the same event or similar event and generate and update schedule information is there.

To achieve the above object, the invention according to claim 1 is characterized in that a message acquisition unit acquires a message via a mail or a bulletin board on the Internet or an intranet, and an event information extraction unit acquires the message. Extracting information representing a specific event from the syntactic structure from the stage, determining the same event or similar event from the extracted event information by the schedule generation unit and automatically generating a schedule, and by the schedule output unit And a step of outputting the generated schedule.
In the invention according to claim 2, the information processing method further includes a step of filtering the generated schedule corresponding to the registered profile or keyword when the user registers a profile or keyword. It is characterized by comprising.

The invention according to claim 3 is characterized in that the message acquisition step accesses all the messages on the bulletin board registered in advance or on the designated bulletin board and acquires the message while maintaining the inter-message relationship.
Further, in the invention according to claim 4, the automatic schedule generation stage uses a knowledge dictionary to identify relationships between elements having the same attribute and different notations, and to determine whether there is an overlapping relationship between the elements. It is characterized by doing.
Further, the invention according to claim 5 is characterized in that the automatic schedule generation step associates a message attachment document or a link document with a schedule.
The invention according to claim 6 is characterized in that the schedule output step automatically converts the schedule information into a standard data format that can be interpreted by a computer.

The invention according to claim 7 is characterized in that in the schedule output stage, time series output, output per place, output per subject, output per person, and output per action are possible.
The information processing apparatus according to the present invention includes a message acquisition unit that acquires a message via a mail or a bulletin board on the Internet or an intranet, and a specific structure using a syntactic structure from the acquired message. An event information extraction unit that extracts information representing an event; a schedule generation unit that automatically generates a schedule by determining the same event or similar event from the extracted event information; a schedule output unit that outputs the generated schedule; It is characterized by comprising.
Further, in the invention according to claim 9, when the user further registers a profile or a keyword, the information processing apparatus filters the generated schedule corresponding to the registered profile or keyword by the filter unit. Characterized in that it comprises steps.

The invention according to claim 10 is characterized in that the message acquisition unit accesses all the messages on the bulletin board registered in advance or on the designated bulletin board and acquires the message while maintaining the inter-message relationship.
In the invention according to claim 11, the schedule generation unit uses a knowledge dictionary to identify a relationship between event elements having the same attribute and different notations, and determines whether there is an overlapping relationship between the elements. It is characterized by that.
The invention according to claim 12 is characterized in that the schedule generation unit associates a message attachment document or a link document with a schedule.

The invention according to claim 13 is characterized in that the schedule output unit automatically converts the schedule information into a standard data format that can be interpreted by a computer.
Furthermore, the invention according to claim 14 is characterized in that the schedule output unit is capable of outputting time-series output, output for each place, output for each subject, output for each person, and output for each action.
The invention according to claim 15 is a program in which a message acquisition unit acquires a message via a mail or a bulletin board on the Internet or an intranet, and an event information extraction unit acquires a syntactic structure from the acquired message. Using the step of extracting information representing a specific event, the step of determining the same event or similar event from the extracted event information by the schedule generation unit and automatically generating the schedule, and the schedule generated by the schedule output unit And outputting.
According to a sixteenth aspect of the present invention, there is provided a recording medium on which the program according to the fifteenth aspect is recorded.

According to the present invention, automatic schedules can be extracted and registered from messages sent daily by electronic communication means without using complicated patterning and costly pattern dictionaries. Collaboration is promoted.
In addition, according to the present invention, for example, a schedule according to a user's application such as a department schedule or a personal schedule can be automatically distributed, and the user only needs to refer to a minimum schedule necessary for the application. User effort is reduced.
Further, according to the present invention, since the relationship between message documents can be acquired, it becomes possible to complement event elements using response relationships and time series, and to determine duplicate events and similar events. High schedule generation can be realized.

Further, according to the present invention, since the relationship between event elements having different notations and the detail level of elements can be specified by the knowledge dictionary, it is possible to determine duplicate events and similar events, and to generate a highly accurate schedule. realizable.
In addition, according to the present invention, since a document attached to an original message document or a linked document can be associated with a schedule, when the user refers to the schedule, the mail software is started or the bulletin board is viewed. Thus, it is possible to open and read the attached document or the linked document without searching for the original message, and the convenience for the user is greatly improved.

Further, according to the present invention, when outputting a schedule, the schedule data can be automatically converted into a common format defined in RFC such as icalender format or Vcal format without user intervention. PIM software that can interpret the format can exchange schedule information, and the same schedule can be managed by a plurality of tools owned by the user, and cooperation between electronic communication and PIM software is promoted.
In addition, according to the present invention, when the user refers to the schedule on the display means, a display format other than time-series display is possible. Therefore, it is possible to refer to the schedule using the element focused on by the user as a key, which greatly improves convenience. To improve.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.
(Example)
FIG. 1 is a configuration example of a computer that realizes an information processing apparatus according to an embodiment of the present invention.
As shown in FIG. 1, the information processing apparatus includes a keyboard 1 as an input means for instructing registration of a processing target document, a user profile, and the like, communication for receiving a signal from the outside and transmitting a signal from the apparatus. Communication I / O interface 2 as a means, CPU 3 for centrally executing processing in this apparatus, memory (both volatile RAM and nonvolatile ROM can be assumed) 4, hard disk 5 as storage means, output means The display 6 and the printer 7 are connected to each other via an internal bus 8.
The communication I / O interface 2 can be a modem or a terminal adapter, and can receive data from a server connected to an intranet or the Internet via a communication line. The CPU 3 executes processing according to the program recorded in the memory 4.

FIG. 2 is an explanatory diagram of a functional configuration of the information processing apparatus according to the embodiment of the present invention.
As shown in FIG. 2, the message acquisition unit 10 is stored in a system having a document database that can be accessed using a network, such as a mail server, a chat server, a bulletin board server, or an RSS distribution server via the Internet or an intranet. Access message documents.
The event information extraction unit 12 reads a message document and extracts all event information that can be acquired in the message. The schedule generation unit 13 determines duplication or similarity between event information and generates a schedule. Here, when a new message is acquired and a new event is added, a duplicate relationship or similar relationship between the new event and the existing event is judged, and if there is no duplicate information, a new schedule is generated, and there is duplicate information. If there is new information, update the existing schedule. If the event is similar to the existing schedule, a similar schedule is generated.

When the user wants to classify the schedule, the input unit 14 receives the user's input, and acquires schedule classification criteria such as keywords and profiles. The filter unit 15 acquires the schedule list and distributes it to the classification schedule list based on the schedule classification standard.
The schedule output unit 16 outputs the entire schedule list or the classified schedule list. In addition, the schedule output unit 16 converts the schedule information into a common format format that can be interpreted by a computer such as icalender. Moreover, the schedule output part 16 delivers update information to a user, when schedule information is updated.

FIG. 3 is a conceptual diagram of processing of the information processing apparatus according to the embodiment of the present invention.
As shown in FIG. 3, when a message is acquired, one or more event information is extracted from the message. When the event information overlaps with the event information of the existing schedule, it is integrated into the existing schedule. If the new event information has new information, the existing schedule is updated. If the event information partially overlaps with the event information of the existing schedule, a similar event is generated. If event information does not overlap with existing schedule event information, a new schedule is generated. If there is a user profile or keyword information, the schedule is filtered based on the information and the schedule list is classified into an arbitrary number.

4 and 10 are explanatory diagrams showing examples of message documents to be processed in one embodiment of the present invention.
Message documents A, B, and C in FIG. 4 are examples of electronic mail documents. The message document B is a reply to the message document A, and the message document C has a document relationship that is a reply to the message document B. Message documents D, E, and F in FIG. 10 are examples of electronic bulletin board documents. The message document E is an opinion / suggestion for the message document D, and the message document F has an inter-document relationship that is an opinion / suggestion for the message document E. When a message document is acquired and read, a message ID that holds the inter-document relationship is assigned.

FIG. 5 is a flowchart of event information extraction processing in the information processing apparatus according to the embodiment of the present invention.
In FIG. 5, the event extraction process determines whether or not there is a target message (step 100), reads the message (step 101), and obtains the syntactic structure of the message text (step 102). An existing natural language analysis process is used for syntactic structure acquisition. At this time, if there is a part that seems to be a quoted part in the text of the message, the part is treated as not being extracted from the specific information set. When the syntactic structure is obtained, information representing the event is extracted (step 103).

  For this event information extraction, for example, 4W1H (When = When, Where = Where) by the method of “information extraction method, information extraction apparatus, and recording medium on which information extraction program is recorded” shown in our related technology (FN200511750) as an example. -Who = who-, What = what-, How = (how and how)) + event information consisting of each element of the predicate is acquired. When event information is extracted from the message (step 104), an event ID is added and registered in the event list together with the event element, message ID, and event ID (steps 105 and 106). If there is other event information in the message, the event information extraction is repeated until there is no event information (step 107). If event information is not acquired, an error is displayed (step 108).

FIG. 6 is an explanatory diagram showing an example of event information extraction in the information processing apparatus according to the embodiment of the present invention, and shows an example in which event information is extracted from the message document B in FIG.
As shown in FIG. 6, the text part of the message document is subjected to natural language analysis, the text is divided into words and clauses, dependency relationships are obtained, and a syntactic structure is obtained. Based on the syntactic structure, 4W1H + predicate information representing an event is extracted. Each event information includes a message ID, an event ID, an event attribute, and an event element.

FIG. 7 is a flowchart of schedule generation processing in the information processing apparatus according to the embodiment of the present invention.
In FIG. 7, when an event is registered (step 200), the message ID, event ID, event attribute, and event element are read and placed in the buffer (step 201). For each event attribute, matching processing is performed between each element and each element in the existing event list (step 202). If there is a duplicate element, duplicate information is added to the target element (steps 203 and 204).
Next, the relationship between elements having the same event attribute and different notations is specified using the knowledge dictionary (steps 205 and 206). If there is an overlap between elements by referring to the knowledge dictionary, the overlap information is given to the target element (steps 207 and 208).

Next, an event element duplication degree determination is performed to determine how many duplicate elements exist between event elements (step 209). In this case, the number of duplicate elements among all event elements is quantified, and a threshold value is set, for example, by the ratio of duplicate elements. If there is an overlap greater than or equal to the threshold, the event is regarded as the same event, a new element is added to the corresponding existing schedule, the schedule is updated, and update information is added to the schedule (steps 210, 211, 212, and 213).
If it is less than the threshold value but there is an overlap, it is regarded as a related schedule, a similar schedule is generated, similar information is given, and a similar schedule ID is acquired (steps 215 and 216). If it is equal to or less than the threshold and there is no overlap, it is regarded as a new schedule, a new schedule is generated, and a schedule ID is acquired (step 217). If there is a schedule similar to the new schedule, it is added to the schedule list (step 218).

  If there are more registration events, this process is repeated until there are no more registration events (step 214). For example, when the schedule of a meeting is adjusted by e-mail, the event is automatically updated and the schedule information is updated as the e-mail is exchanged. When mail exchange is interrupted, the schedule is stored in a static state. In addition, if there is a meeting schedule on the electronic bulletin board, and the minutes are uploaded to the same message tree on the same bulletin board after the meeting ends, it is automatically determined whether the event of the message is the same as the target schedule. Judgment is made and the minutes are associated with each other and the target schedule is updated. Duplicate information and similar information are expressed by, for example, flags and specific codes.

FIG. 8 is an explanatory diagram of a description example of the knowledge dictionary in the information processing apparatus according to the embodiment of the present invention.
As shown in FIG. 8, for each attribute of the event, a pair of event element notation is written on the left side and the right side. In this example, it is written in a solid and easy-to-understand manner. However, the description may be in the form of grammatical information expressed in regular expressions, and in the case of a system implementation, rules that write down this knowledge in the form of program code are used. Is appropriate. When the notation of the event element of an attribute with a registered event matches the element on the left side, 1 is subtracted from the level of detail. If it matches the element on the right side, add 1 to the level of detail. In the comparison with the event element of the existing schedule, it is used for specifying the relationship between the different notation elements, and an element with a high degree of detail is used as an element for schedule generation.

FIG. 9 is an explanatory diagram of a duplicate event example and a schedule generation example in the information processing apparatus according to an embodiment of the present invention.
FIG. 11 is an explanatory diagram of a schedule update example in the information processing apparatus according to the embodiment of this invention.
12 is an explanatory diagram of an example in which the element of the schedule ID 001 of the schedule list in FIG. 11 in the information processing apparatus according to the embodiment of the present invention is converted into the icalender format. After the schedule information is generated / registered by the information processing of the present invention, the schedule information can be converted into, for example, an icalender format and stored at the time of output.
FIG. 13 is an explanatory diagram of a schedule display example in the information processing apparatus according to the embodiment of this invention.

Hereinafter, specific examples of the information processing apparatus according to the present invention will be described with reference to FIGS. 4, 6, and 8 to 13.
Specific Example 1: Schedule Generation / Display For example, when message groups A, B, and C in FIG. 4 are registered in the information processing apparatus according to the embodiment of the present invention, the message acquisition unit First, header information of a message document is acquired, and since all header parts have “X-Mailer: OO”, it is specified that these documents are a message document group acquired from the electronic mail system.

  Next, when the message document A is the target document and the message document B is the next document, the Message-Id of the next document is newer than the Message-Id of the target document, and the Message-Id of the next document In-Reply-To is Message-Id of the target document, the date of the next document is newer than the date of the target document, the subject of the next document has the same character string as the subject of the target document, and the subject of the next document is prefixed with Re: Thus, the message ID of the message document A is assigned 001, and the message ID of the message document B is assigned 002.

Next, when the message document B is the target document and the message document C is the next document, the Message-Id of the next document is newer than the Message-Id of the target document, and the Message-Id of the next document In-Reply-To is Message-Id of the target document, the date of the next document is newer than the date of the target document, the subject of the next document has the same character string as the subject of the target document, and the subject of the next document is prefixed with Re: Since the message ID of the message document B is 002, 003 obtained by adding 1 to the message ID of B is added to the message ID of the message document C.
By the above processing, the message document group in FIG. 4 is an electronic mail document group, the message document A is a mail original document, the message document B is a reply to the message document A, and the message document C is a reply to the message document B. Can be specified.

When the message acquisition unit 10 determines the relationship between the messages, the message acquisition unit 10 stores the target message document together with the message ID in the message list.
The event information extraction unit 12 is activated when a new message is added to the message list, and determines whether there is a target message. Since there are target message documents A, B, and C, the message document A is first read and event information is extracted. If there is a part that seems to be a quoted part in the text of the message, that part is treated as an event extraction target.

統語構造を取得したら、文末から述語を検索し、述語毎に述語と係り受け関係にあるイベント要素をイベント情報として抽出する。メッセージ文書のイベント情報が抽出されたら、イベントＩＤを付加してイベントリストにメッセージＩＤ、イベントＩＤと共にイベント属性とイベント要素を登録する。イベント属性とは、４Ｗ１Ｈ（Ｗｈｅｎ＝いつ、Ｗｈｅｒｅ＝どこ−で、Ｗｈｏ＝誰−が、Ｗｈａｔ＝何−を、Ｈｏｗ＝（どのくらい・どうやって））＋述語のことである。１つのメッセージ内に、別のイベント情報があれば、イベント情報がなくなるまでイベント情報抽出を繰り返す。この結果、上記統語構造からは、以下の００１〜００３のイベント情報が取得できた。
メッセージ文書Ｂ＜メッセージＩＤ ００２＞
００１述語｛佐藤・です・。｝
００１Ｈｏｗ｛第・一・開発・部｝
００２述語｛いい・です・。｝
００２０Ｗｈｅｎ｛来月・７日｝
００２２Ｗｈｅｎ｛午前・中｝
００３述語｛どこ・です・か・？｝
００３Ｈｏｗ｛場所｝
このようなイベント情報抽出手続きを、メッセージ文書Ａとメッセージ文書Ｃに対しても同様に実施し、各メッセージから、図９に示すようなイベント情報を抽出できた。 The event information extraction will be described using an example of event information extraction from the message document B in FIG. First, the syntactic structure of the text part of the message document B as follows is acquired.

When the syntactic structure is acquired, a predicate is searched from the end of the sentence, and event elements having a dependency relationship with the predicate are extracted as event information for each predicate. When the event information of the message document is extracted, an event ID is added and an event attribute and an event element are registered together with the message ID and event ID in the event list. The event attribute is 4W1H (When = when, Where = where−, Who = whose−, What = what−, How = (how / how)) + predicate. If there is other event information in one message, the event information extraction is repeated until there is no event information. As a result, the following event information of 001 to 003 could be acquired from the syntactic structure.
Message document B <Message ID 002>
001 predicate {is Sato ... }
001 How {No. 1, Development, Department}
002 predicate {good. }
0020When {Next month, 7th}
0022When {Morning / Middle}
003 predicate {where is it? }
003How {location}
Such an event information extraction procedure was similarly performed for the message document A and the message document C, and event information as shown in FIG. 9 could be extracted from each message.

The schedule generation unit 13 is activated when a new event is added to the event list, and determines whether there is a target event. Next, when an event is registered, the message ID, event ID, event attribute, and event element are read and placed in the buffer. Then, it is checked whether there is a match with an existing event list element. If there is a match, it is considered that there is an overlap of elements, and duplicate information is added to the duplicate event element.
For example, when a new event of the message document A is newly added, the message ID, event ID, event attribute, and event element are read in order from the top of the four events of the message document A in FIG. Then, when checking whether there is a match with an element of the existing event list, there is no match, so next, it is checked whether there is information having the same event attribute but different element notation. Since there is no information with the same event attribute and different element notation, a new schedule based on this event is generated, and after acquiring the schedule ID, it is registered and stored in the schedule list.

Next, when a new event of the message document B is added, the message ID, event ID, event attribute, and event element are read in order from the top of the three events of the message document B in FIG. Then, with reference to the presence / absence of matching with the elements of the existing event list and the knowledge dictionary of FIG.
Since the event attribute When “next month” of the event ID 0020 and the event attribute When “morning” of the event ID 0021 coincide with the left side element of the knowledge dictionary, the elements are changed to the right side elements “September” “{0 | 1 | 2 | 3 | 4”. | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12} ", and subtract 1 from each detail level. There is no overlap with existing event elements or information with the same event attribute but different element notation, but since there is a relationship between message document A and message document, it is regarded as the same event and the schedule information is updated by adding new elements. The update information is added to the schedule information.

Next, when a new event of the message document C is added, the message ID, event ID, event attribute, and event element are read in order from the top of the four events of the message document C in FIG. Then, it is checked whether there is a match with an existing event list element and whether the event element has a different notation. The element “7 days” of the event attribute “When”, the element “setting” of the event attribute predicate, the element “conference” of the event attribute What, and the elements “Yamada” and “Sato” of the event attribute Who match, giving duplicate information.
Next, it is checked whether there is information with the same event attribute but different element notation. Since the event attribute When “10 o'clock to 12 o'clock” of the event ID 0031 matches the right side of the knowledge dictionary, the detail level of the element is incremented by one. Since this event ID 0031 matches the event attribute When “{0 | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12}" of the event ID 0021, duplicate information is given. To do. Furthermore, the event attribute Where “first conference room” is set as a new event element. Since the event ID 0021 and the event ID 0031 have a high degree of duplication, they are regarded as the same event, and an event attribute element having a high degree of detail is adopted as the optimum event element, the schedule is updated, and update information is given.

  The schedule output unit 16 displays the entire schedule list or the substance of the filtered schedule list on a display device such as a display. FIG. 13 shows a schedule list displayed on the display. Since the schedule is composed of event attribute elements of 4W1H + predicate, it is possible to design the display to be switched based on arbitrary criteria of a plurality of users, such as not only time-series display but also display for each location. The display may be switched using a tab or the like. Further, the schedule output unit 16 can convert the schedule information into a computer-readable interpretable common format such as icalender, and output the same to a schedule tool of PIM software that complies with this format. FIG. 12 shows an example in which the schedule data of the schedule ID 001 in the schedule list of FIG. Further, the schedule output unit 16 can be designed to distribute the update information to the user when the schedule information is updated.

Specific Example 2: Schedule Filtering / Display Next, a specific example of schedule filtering will be described. The process until the event information is automatically extracted from the message document and the event information is automatically edited to create the schedule list is the same as in the first specific example. When displaying the schedule list in advance or before display, the user can enter keywords such as “Yu-san Co., Ltd.” or “Launch”, the user ’s age / sex / hobby / work theme, etc. And the like, and the schedule information is filtered using the information to create a keyword-specific schedule list and a user-specific schedule list.

For example, the keyword “Mr. XX Co., Ltd.” is registered in the element of the event attribute Who, and when a new event is registered, it is checked whether it matches the element of the new event attribute Who. It is output as a schedule for XX Corporation. Similarly, if “sale” is registered in the element of the event attribute predicate, when a new event is registered, it is checked whether it matches the element of the new event attribute predicate. Output as.
At this time, the substance of the schedule for “S Co., Ltd.” and the “release” schedule is stored in the original schedule list, and the schedule list for “S Co., Ltd.” and the “sale” schedule list have schedule IDs. The list is effective without affecting the processing speed. Similarly, when the user registers a work theme in the element of the event attribute What, a schedule related to the theme is output.

Specific Example 3: Attached Document Linked Document Next, a specific example of associating an attached document or a linked document with a schedule will be described with reference to FIGS. For example, in a specific electronic bulletin board, the message document D and the message document E in FIG. The event information extraction unit 12 extracts the following four events from the message document D.
Message document D <message ID 101>
001 predicate {Yamada. }
002 predicate {-}
0020When {next month}
0021When {Thursday, September 22}
0022When {From 13:00}
002What {meeting}
003 predicate {please keep it}
003What {planned}
003Who {Theme Member}
004 predicate {Thank you, please. }

The schedule generation unit 13 performs a duplication check and a similarity check with the existing event information, acquires the schedule ID 002 as a new schedule, and registers it in the schedule list. Next, the event information extraction unit 12 extracts the following three events from the message document E.
Message document E <message ID 102>
001 predicate {is Sato ... }
002 predicate {-}
002Where {First Conference Room}
002What {meeting / location}
003 predicate. }
003When {from 13:00 to 15:00}

The schedule generation unit 13 performs a duplication check and a similarity check with the existing event information, and since there are more than a certain number of duplicate event elements, the schedule generation unit 13 determines that the schedule ID002 is duplicated and updates the schedule ID002 by adding a new event element. . The schedule ID 002 in the schedule list of FIG. 11 is in this state.
When the message document F shown in FIG. 10 is uploaded to the same message tree on the same bulletin board after the meeting is actually finished, the event information extraction unit extracts from the message document F the following three events and one Extract the attached document.
Message document F <message ID 103>
001 predicate {Yamada. }
002 predicate {create}
002What {meetings / minutes}
003 predicate {please confirm / please. }
003Who {Theme Member}
004 Attachment:

  The schedule generation unit 13 performs a duplication check and a similarity check with the existing event information, determines that the event is the same as the schedule of the schedule ID 002, adds a new event element, automatically updates the schedule ID 002, and updates the update information. Give. The schedule ID 002 (update) after schedule update in FIG. 11 is in this state. This schedule ID 002 (update) is associated with a message document F and an attached document. In the present invention, since the schedule output unit 16 can distribute the update information of the schedule to the user, the user can receive the update information and refer to the minutes even if the meeting is over and the schedule is not reviewed. It becomes.

Specific Example 4: Schedule Information Conversion Processing FIG. 12 is an example in which the schedule information with the schedule ID 001 generated in FIGS. 4, 6, and 9 is automatically converted into an icalender format, stored, and output. For example, it is assumed that the PIM system incorporating the information processing apparatus according to the present invention is activated and the message documents A to C in FIG. 4 are registered. At this time, at the same time as document registration, the PIM system incorporating the information processing apparatus of the present invention stores the registered message document, and performs processing similar to the event information extraction processing desired in the previous embodiment in order from the first message document. do. FIG. 6 shows the event information extraction result from the message document B. As shown in FIG. 11, event information of each message document is obtained and stored.

Next, for example, a URI defining a vocabulary having 4W1H + predicate information as a property element, in this embodiment, http://xxxx.com//RDFCal1.0//EN is prepared in advance and used together with an existing vocabulary. To do. (See the icalender format conversion example in FIG. 12) If there is an existing vocabulary that matches the target event element, it is not necessary to prepare a newly defined vocabulary.
First, what information “meeting” is described as a Summary element. Next, When information “September”, “Seventh”, and “10:00 to 12:00” are divided into a start time and an end time, and are described as a DTSTART element and a DTEND element. Next, the Where information “first conference room” is described as a LOCATION element.
These are stored, and if there is an output instruction, an output process is performed. As an example, FIG. 12 shows an example of icalender format conversion of schedule information.

As described above, according to the embodiment of the present invention, since there is no manual process for schedule registration and schedule editing, it is not necessary to request the user for registration editing time and effort. In addition, since the syntactic structure is acquired using the advanced natural language processing technology and the event information is automatically extracted, it is not necessary to construct a message structure or language pattern suitable for the message in advance.
In addition, when outputting a schedule, the schedule data can be converted into a common format defined by RFC, such as the icalender format or Vcal format, so that schedule information can be exchanged between PIM software that can interpret this format. The same schedule can be managed by a plurality of tools possessed by the user. In addition, it has the feature that event information expressed across multiple messages and the same event information mentioned in different messages can be handled as the same schedule, and is equally divided into the number of large numbers of messages. If a new message is acquired, it is not treated as a new schedule, but it is determined whether it is the same as an event of a registered schedule. The registered schedule can be updated.
Therefore, by using the method, apparatus, and program according to the present invention, automatic schedule generation, display, and distribution can be realized easily and accurately from messages sent daily by electronic communication means.

It is a figure which shows the structural example of the computer which implement | achieves the information processing apparatus which is one Embodiment of this invention. It is explanatory drawing of the function structure of the information processing apparatus which is one Embodiment of this invention. It is a process conceptual diagram of the information processing apparatus which is one Embodiment of this invention. It is explanatory drawing which shows the example of the message document used as the process target in one Embodiment of this invention. It is a flowchart of the event information extraction process in the information processing apparatus which is one Embodiment of this invention. It is explanatory drawing which shows the example of event information extraction in the information processing apparatus which is one Embodiment of this invention. It is a flowchart of the schedule production | generation process in the information processing apparatus which is one Embodiment of this invention. It is explanatory drawing of the example of a description of the knowledge dictionary in the information processing apparatus which is one Embodiment of this invention. It is explanatory drawing of the example of a duplication event in the information processing apparatus which is one Embodiment of this invention, and a schedule production | generation example. It is explanatory drawing which shows the example of the message document used as the process target in one Embodiment of this invention. It is explanatory drawing of the example of a schedule update in the information processing apparatus which is one Embodiment of this invention. It is explanatory drawing of the example which converted the element of schedule ID001 of the schedule list | wrist of FIG. 11 in the information processing apparatus which is one Embodiment of this invention into icalender form. It is explanatory drawing of the example of a schedule display in the information processing apparatus which is one Embodiment of this invention.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... Keyboard, 2 ... Interface, 3 ... CPU, 4 ... Memory, 5 ... Hard disk, 6 ... Display, 7 ... Printer, 8 ... Internal bus, 10 ... Message acquisition part, 12 ... Event information extraction part, 13 ... Schedule generation 14, input unit 15, filter unit

Claims (16)

A step of acquiring a message via a mail or a bulletin board on the Internet or an intranet by a message acquisition unit;
Extracting information representing a specific event from the acquired message by the event information extraction unit using a syntactic structure;
A step of automatically generating a schedule by determining the same event or similar event from the extracted event information by the schedule generation unit,
A step of outputting the generated schedule by a schedule output unit.   The information processing method further comprises a step of filtering the generated schedule corresponding to the registered profile or keyword by the filter unit when the user registers a profile or keyword. The information processing method according to 1.   3. The information processing method according to claim 1, wherein the message obtaining step accesses all messages on a bulletin board registered in advance or on a designated bulletin board, and obtains the message while maintaining the inter-message relationship.   The schedule automatic generation step is characterized in that, by using a knowledge dictionary, a relationship between elements having the same attribute and different notations is specified, and the presence or absence of an overlapping relationship between the elements is determined. 3. The information processing method according to 2.   The information processing method according to claim 1, wherein the automatic schedule generation step associates a message attachment document, a link document, or the like with a schedule.   3. The information processing method according to claim 1, wherein the schedule output step automatically converts the schedule information into a standard data format that can be interpreted by a computer.   3. The information processing method according to claim 1, wherein the schedule output step is capable of outputting each of time series output, output for each place, output for each subject, output for each person, and output for each action. A message acquisition unit that acquires messages via mail or bulletin boards on the Internet or intranet,
An event information extraction unit for extracting information representing a specific event from the acquired message using a syntactic structure;
A schedule generation unit for automatically generating a schedule by determining the same event or similar event from the extracted event information;
An information processing apparatus comprising: a schedule output unit that outputs the generated schedule.   The information processing apparatus further includes a step of filtering the generated schedule corresponding to the registered profile or keyword by the filter unit when the user registers a profile or keyword. The information processing apparatus according to 8.   The information processing apparatus according to claim 8 or 9, wherein the message acquisition unit accesses all messages on a previously registered bulletin board or a designated bulletin board, and acquires the message while maintaining the inter-message relationship.   The schedule generation unit identifies a relationship between elements that are event elements having the same attribute and different notations using a knowledge dictionary, and determines whether or not there is an overlapping relationship between the elements. The information processing apparatus described in 1.   The information processing apparatus according to claim 8, wherein the schedule generation unit associates a message attachment document, a link document, or the like with a schedule.   The information processing apparatus according to claim 8, wherein the schedule output unit automatically converts the schedule information into a standard data format that can be interpreted by a computer.   The information processing apparatus according to claim 8, wherein the schedule output unit is capable of outputting time-series output, output for each place, output for each subject, output for each person, and output for each action. A step of acquiring a message via a mail or a bulletin board on the Internet or an intranet by a message acquisition unit;
Extracting information representing a specific event from the acquired message by the event information extraction unit using a syntactic structure;
A step of automatically generating a schedule by determining the same event or similar event from the extracted event information by the schedule generation unit,
Outputting the generated schedule by a schedule output unit.   A recording medium on which the program according to claim 15 is recorded.

Images