JP2005182400A - Natural language processor, processing method and program, and portable terminal loaded with natural language processor - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a natural language processor, processing method and a program and a portable terminal loaded with the natural language processor for easily browsing opposite party information in an arranged form by automatically extracting from the electronic mail the information (attributes and conditions, etc.) of an opposite party with whom electronic mail is exchanged, and storing the information as a database even for a small-sized device with limited hardware resources. <P>SOLUTION: The natural language processor is provided with: a morpheme analysis part for performing a the processing of dividing a sentence included in received electronic mail into word units; a relation analysis part for performing the processing of extracting the opposite party information on the basis of the analysis result of the morpheme analysis part; a dictionary for morpheme analysis/relation analysis storing necessary minimum words and rules for the relation analysis to be required when performing registration in an opposite party information database as the opposite party information and to be referred to when the morpheme analysis part and the relation analysis part perform analysis/extraction processing; and a control part. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to a natural language processing apparatus, method and program, and a portable terminal equipped with a natural language processing apparatus, and in particular, information (attributes, status, etc.) of a partner with whom an electronic mail is exchanged is automatically transmitted from the electronic mail. The present invention relates to a natural language processing apparatus, method, and program that can be extracted and stored as a database, and a portable terminal equipped with the natural language processing apparatus.

  In recent years, e-mail using personal computers and portable terminals has become widespread, and it has become desirable to provide a user interface that can easily obtain information (attributes, status, etc.) of the partner with whom e-mail is exchanged. ing.

  However, in order to obtain information on the other party with whom an email is exchanged based on the email, it is premised that a huge amount of past emails must be saved, It is necessary to search and confirm necessary information one by one from the e-mail.

  The e-mail contains a lot of useful information for the sender / receiver or between the parties concerned. However, when trying to obtain the useful information based on the e-mail, as described above, it is enormous. It is assumed that a large amount of past e-mails is stored, and it is necessary to search and confirm useful information to be obtained one by one from the past e-mails.

  In order to be able to efficiently check a large amount of e-mails later, it has been proposed as a solution to store and store these e-mails in a database by theme (for example, (See Patent Document 1).

  In the “information management device, communication terminal, information management system, information management method, recording medium, and program” described in Patent Document 1 described above, business documents (e-mail) exchanged between specific persons are It is described that the database is stored separately in each database, and this allows business documents stored in the database to be shared among the parties involved in the business task.

  There has also been proposed an e-mail apparatus that automatically performs an operation of searching and checking useful information one by one from an e-mail without causing an increase in work load (see, for example, Patent Document 2). .

  In the above-mentioned “recording medium on which an electronic mail apparatus and an electronic mail program are recorded” described in Patent Document 2, the CPU of the electronic mail apparatus extracts a character string described in a predetermined portion from the received electronic mail. A technique is described in which the extracted character string is registered in a predetermined item corresponding to the classification of address book data managed by, for example, an address book program. As a result, necessary information is extracted and registered from a part described in a predetermined part (e.g., signature) of received e-mail according to a plurality of items classified and stored by address book data. In addition, the contents of the e-mail can be effectively used without special editing work or the like.

  Furthermore, when extracting useful information (attribute information) from e-mails, we also propose an e-mail system that analyzes e-mail texts to obtain word segmentation and word information, and extracts attribute information from texts using a dictionary. (For example, see Patent Document 3).

  The “e-mail system” described in Patent Document 3 described above uses a dictionary to analyze a text in an email to obtain word division and word information, and to use this to analyze a header and An e-mail system is described that includes an attribute information extraction unit that extracts attribute information from a body part, and a header update unit that adds the attribute information as extension information to the header part of the e-mail. With this system, it is possible to extract and display e-mail attribute information.

JP 2003-016019 A (page 5-12, FIG. 1-30) JP 2000-172587 A (page 3-6, FIGS. 1-4) Japanese Patent Laid-Open No. 10-078923 (page 3-6, FIG. 1-12)

  In the above-described conventional information management apparatus and e-mail system, it is necessary to store and store the received e-mail as a database by theme, or to extract useful information such as address book data from the received e-mail. It is possible to register and use in locations, analyze email text and extract email attribute information, etc., and automatically collect useful information from a huge amount of email It is possible to extract and save.

  However, if it is possible to provide a user interface that automatically extracts and saves information such as the attributes and status of the person with whom the email is exchanged from those emails, it has been received in the past The trouble of manually picking up and storing necessary information from the e-mail is reduced, and the convenience for the user is improved.

  In addition, if e-mail text analysis processing can be realized by a simpler method than the e-mail system described in Patent Document 3, for example, it is a small device with limited hardware resources, such as portable terminals that have become popular in recent years. However, it can be easily mounted, and the convenience for the user is further improved.

  The present invention has been made in order to improve the above-described circumstances. Therefore, the object of the present invention is to automatically extract information (attributes, status, etc.) of the partner with whom an email is exchanged from the email. Can be stored as a database, and even small devices with limited hardware resources can be easily installed, so that partner information can be easily viewed in a unified form. Another object of the present invention is to provide a natural language processing apparatus, method and program, and a portable terminal equipped with the natural language processing apparatus.

The natural language processing apparatus of the present invention is a natural language processing apparatus that extracts partner information including attributes and conditions of a partner with whom an email is exchanged from a received email and registers it in the partner information database.
The natural language processing device includes a morpheme analyzer that performs processing to divide a sentence included in a received email into words.
A relation analysis unit that performs processing to extract partner information based on the analysis result of the morphological analysis unit;
A morpheme that is stored as the partner information in the partner information database and stores the minimum necessary words and relation analysis rules and is referred to when the morpheme analysis unit and the relationship analysis unit perform analysis / extraction processing. A dictionary for analysis and relation analysis,
Control for causing the morpheme analysis unit to analyze the email body of the received email, control for causing the relationship analysis unit to perform relationship analysis based on the analysis result of the morpheme analysis unit, and execution of the partner information extraction process, and the relationship analysis A control unit for performing the control of adding the e-mail address and time stamp information of the e-mail header information of the received e-mail to the result of the extraction process of the unit and registering the information in the counterpart information database;
It is characterized by providing.

  The natural language processing device further includes a partner information history management unit that manages partner information as a history, and the control unit obtains the result of the extraction process of the relation analysis unit, and the partner information history The management unit confirms whether the partner information of the same category as the result of the extraction process has already been registered in the partner information database. If the partner information of the same category has already been registered, the existing partner information is extracted. It is characterized in that control is performed to replace the result of processing for registration.

  Further, when the natural language processing device obtains the browsing request information of the partner information from the input device, it searches the partner information database, extracts the partner information that matches the browsing request information, and outputs it to the output device. It is characterized by that.

The natural language processing method of the present invention is a natural language processing apparatus that extracts partner information including attributes and conditions of a partner with whom an email is exchanged from a received email and registers it in the partner information database.
The natural language processing device includes a morpheme analyzer that performs processing to divide a sentence included in a received email into words.
A relation analysis unit that performs processing to extract partner information based on the analysis result of the morphological analysis unit;
A morpheme that is stored as the partner information in the partner information database and stores the minimum necessary words and relation analysis rules and is referred to when the morpheme analysis unit and the relationship analysis unit perform analysis / extraction processing. And a dictionary for analysis and relation analysis,
The natural language processing device causes the morphological analysis unit to analyze the email body of the received e-mail, and causes the relationship analysis unit to analyze the relationship from the analysis result of the morpheme analysis unit, and executes partner information extraction processing And a step of adding a mail address and time stamp information of the e-mail header information of the received e-mail to the result of the extraction process of the relation analysis unit and registering the result in the partner information database. Features.

  The natural language processing device further includes a partner information history management unit that manages partner information as a history, and when the natural language processing device obtains a result of the extraction process of the relation analysis unit, Let the information history management unit check whether the partner information of the same category as the result of the extraction process has already been registered in the partner information database, and if the partner information of the same category has already been registered, the existing partner information is The result of registration is replaced with the result of the extraction process.

  Further, when the natural language processing device obtains the browsing request information of the partner information from the input device, the natural language processing device searches the partner information database, extracts the partner information that matches the browsing request information, and outputs the partner information to the output device. It is characterized by performing.

The program of the present invention includes a morpheme analyzer that performs a process of dividing a sentence included in a received e-mail into word units;
A relational analysis unit that performs processing to extract partner information consisting of attributes and status of a partner who is exchanging e-mails based on the analysis result of the morphological analysis unit;
Morphological analysis that is stored in the partner information database as partner information and stores the minimum necessary words and relation analysis rules, and is referred to when the morpheme analysis unit and the relationship analysis unit perform analysis / extraction processing A relational analysis dictionary, and a computer of a natural language processing device that extracts partner information from received e-mail and registers it in the partner information database;
A process for causing the morphological analysis unit to analyze the body of the email of the received e-mail, a process for causing the relational analysis unit to perform a relational analysis based on an analysis result of the morpheme analysis part, and a process for executing the partner information extraction process. A process of adding the mail address and time stamp information of the e-mail header information of the received e-mail to the result of the part extraction process and registering the result in the partner information database.

  The natural language processing apparatus further includes a partner information history management unit that manages partner information as a history, and the program receives the result of the extraction process from the relation analysis unit when the computer obtains a result of the extraction process. Let the information history management unit check whether the partner information of the same category as the result of the extraction process has already been registered in the partner information database, and if the partner information of the same category has already been registered, the existing partner information is It is characterized in that a process of registering with the result of the extraction process is executed.

  Further, when the program obtains the browsing request information of the partner information from the input device, the program executes a process of searching the partner information database to extract partner information that matches the browsing request information and outputting it to the output device. It is characterized by that.

A mobile terminal equipped with the natural language processing device of the present invention is a natural language processing device that extracts partner information including attributes and conditions of a partner with whom an email is exchanged from a received email and registers it in the partner information database. An on-board mobile device,
The natural language processing device includes a morpheme analyzer that performs processing to divide a sentence included in a received email into words.
A relation analysis unit that performs processing to extract partner information based on the analysis result of the morphological analysis unit;
A morpheme that is stored as the partner information in the partner information database and stores the minimum necessary words and relation analysis rules and is referred to when the morpheme analysis unit and the relationship analysis unit perform analysis / extraction processing. A dictionary for analysis and relation analysis,
Control for causing the morpheme analysis unit to analyze the email body of the received email, control for causing the relationship analysis unit to perform relationship analysis based on the analysis result of the morpheme analysis unit, and execution of the partner information extraction process, and the relationship analysis A control unit for performing the control of adding the e-mail address and time stamp information of the e-mail header information of the received e-mail to the result of the extraction process of the unit and registering the information in the counterpart information database;
It is characterized by providing.

  The natural language processing device further includes a partner information history management unit that manages partner information as a history, and the control unit obtains the result of the extraction process of the relation analysis unit, and the partner information history The management unit confirms whether the partner information of the same category as the result of the extraction process has already been registered in the partner information database. If the partner information of the same category has already been registered, the existing partner information is extracted. It is characterized in that control is performed to replace the result of processing for registration.

  Further, when the natural language processing device obtains the browsing request information of the partner information from the input device, it searches the partner information database, extracts the partner information that matches the browsing request information, and outputs it to the output device. It is characterized by that.

  The natural language processing device, method and program of the present invention, and a portable terminal equipped with the natural language processing device can automatically extract the partner information from the received e-mail and store the partner information in the partner information database. It is possible to easily view the partner information in a collective form without storing and retrieving a large amount of past emails, and the user can manually read the partner information from the received email. This eliminates the need for the work of extracting, so that the user's labor can be reduced and the convenience of the user is improved.

  Further, in the present invention, a relational analysis process is added to the morphological analysis process. By limiting the morphological analysis process to a degree necessary for the relational analysis process, the morphological analysis / relationship analysis dictionary includes a relational analysis. By registering only the minimum words and rules necessary for processing, a natural language processing device can be easily installed even in a small device such as a portable terminal with limited hardware resources. The effect that the speed of language processing improves can be produced.

  Furthermore, in the present invention, by registering the partner information in the partner information database in time series using the time stamp of the e-mail, there is an effect that the partner information can be browsed in time series. Can do.

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

  FIG. 1 is a block diagram showing an embodiment of a portable terminal equipped with the natural language processing apparatus of the present invention.

  The present embodiment shown in FIG. 1 is applied to a mobile terminal 9 including a mobile phone, a PHS (Personal Handyphone System) terminal, or a PDA (Personal Digital Assistance). 1 shows an embodiment in which a processing apparatus is mounted. The portable terminal 9 is assumed to be equipped with an electronic mail transmission / reception device (not shown).

  In FIG. 1, a portable terminal 9 includes an input device 1 such as a keyboard and a key input device, an output device 2 such as a display device and a printing device, and a partner information extraction processing device that performs processing for extracting partner information from received electronic mail. 3, a received e-mail database 4 that stores and stores received e-mails, and a partner information database 5 in which partner information extracted by the partner information extraction processing device 3 is registered.

  Here, the partner information is information indicating the attribute and status of the sender of the electronic mail, that is, the partner. The other party's attributes are information on characteristics / characteristics such as the other party's address, name, age, date of birth, occupation, telephone number, email address, etc. The situation of the other party means that the other party has traveled This is event information about the recent situation of the other party that can be known from the received e-mail.

  The partner information extraction processing device 3 includes a control unit 31 that controls the operation of the partner information extraction processing device 3 as a whole, a morpheme analysis unit 32 that performs morphological analysis of received e-mails under the control of the control unit 31, and control of the control unit 31. Based on the analysis result of the morpheme analysis unit 32, the relationship analysis unit 33 that analyzes and extracts the partner information, the morpheme analysis / relation analysis dictionary 34 that the morpheme analysis unit 32 and the relationship analysis unit 33 refer to when performing analysis, The storage unit 35 stores and stores the control program of the control unit 31.

  The control unit 31 includes a CPU (Central Processing Unit) and is configured to operate according to a control program stored in the storage unit 35. Note that the processes of the morphological analysis unit 32 and the relationship analysis unit 33 may be configured to operate according to a control program stored in the storage unit 35.

  The control unit 31 controls access to the received email database 4 and the partner information database 5, operation control of the morphological analysis unit 32 and the relationship analysis unit 33, input information control from the input device 1, output information control to the output device 2, I do.

  The morpheme analysis unit 32 uses the morpheme analysis / relationship analysis dictionary 34 to perform morpheme analysis on the received e-mail. The morpheme analysis process is a process of dividing a sentence included in the received electronic mail into words (morphemes).

  The relationship analysis unit 33 uses the analysis result of the morpheme analysis unit 32 and the morpheme analysis / relationship analysis dictionary 34 to perform processing for extracting attributes, status, and other information related to the partner, that is, partner information. The partner information extracted by the relationship analysis unit 33 is classified into a category such as the attribute and status of the partner via the control unit 31 and registered in the partner information database 5.

  The morpheme analysis / relationship analysis dictionary 34 is a dictionary that stores and stores words and relationship analysis rules that are required when the morpheme analysis unit 32 and the relationship analysis unit 33 perform analysis processing. Words necessary for analysis processing are, for example, parts of speech that make up sentences such as nouns, verbs, particles, and auxiliary verbs. Specifically, “I” (pronoun), “ha” (adjunct particle) ), “University student” (noun), “ni” (case particle), “naru” (simplified word in this embodiment), and the like. In addition, “Naru” is usually “Nari” (“Naru”: intransitive verb), “Mashi” (“Masu”: auxiliary verb / special type), “Ta” (auxiliary verb / special type), Although the word is further subdivided, in this embodiment, “Nana” is used as one word in order to simplify the word. As a result, the number of registered words in the morphological analysis / relationship analysis dictionary 34 is reduced, thereby reducing the memory capacity required for the dictionary.

  Also, the relational analysis rules necessary for performing the analysis process indicate the criteria for extracting the partner information. For example, from the sentence “x has become a college student”, “x The rule is that the partner information “= University student” is extracted. This indicates that a word (actor) that applies to x on the left side has a relationship of “x = university student”.

  The relation analysis rule also indicates a standard for registering the extracted partner information in the partner information database 5. For example, the rule “profession: university student” indicates “college student” as the partner information database 5. This indicates the standard of registering as a “profession category”. Further, the rule “I: mail sender” indicates a criterion for registering in the partner information database 5 when the actor is “I”.

  Next, the operation of the present embodiment shown in FIG. 1 will be described with reference to FIG.

  FIG. 2 is a flowchart for explaining the operation of the embodiment of FIG.

  First, triggered by the reception of the email, the control unit 31 of the partner information extraction processing device 3 takes in the email received from the received email database 4 (step A1 in FIG. 2). Of the captured e-mail (consisting of e-mail header information and e-mail body), the e-mail body is passed to the morpheme analyzer 32, and e-mail header information such as e-mail address and time stamp is stored in the partner information database. 5 is used by the control unit 31 to be used when registering to 5.

  The control unit 31 activates the morpheme analysis unit 32. The morpheme analysis unit 32 receives the email text from the control unit 31 and performs morpheme analysis processing using the morpheme analysis / relationship analysis dictionary 34 (step A2). . As described above, the morpheme analysis process is a process of dividing a sentence included in the received e-mail into words (morpheme) units. However, in the present embodiment, it is only necessary to extract a word to be registered as partner information. Therefore, the morphological analysis is performed with an accuracy that does not hinder the analysis processing of the relationship analysis unit 33 to be executed thereafter. Specifically, as described above, “Naru” is handled as one word. The result of the morphological analysis performed by the morphological analysis unit 32 is passed to the control unit 31.

  The control unit 31 to which the result of the morphological analysis is passed next activates the relationship analysis unit 33 and passes the result of the morphological analysis to the relationship analysis unit 33. The relationship analysis unit 33 receives the result of the morphological analysis from the control unit 31, performs the relationship analysis processing using the morphological analysis / relation analysis dictionary 34, and performs the partner information extraction processing (step A3). As described above, the relationship analysis process is a process for extracting attributes and status of the other party, and the morphological analysis / relationship analysis dictionary 34 has a minimum word required for extracting the partner information. And the relation analysis rules are stored, the relation analysis unit 33 can extract the partner information using the result of the morphological analysis and the morphological analysis / relation analysis dictionary 34. The result of the relationship analysis performed by the relationship analysis unit 33 is passed to the control unit 31.

  The control unit 31 registers the result of the relationship analysis, that is, the obtained partner information in the partner information database 5 with information such as a mail address and a time stamp added (step A4). In the partner information database 5, the partner information is classified into categories such as attributes and situations, and the partner information is arranged in time series using a time stamp and added and registered as one record in the database.

  After the above processing, by operating the input device 1 and accessing the partner information database 5, the control unit 31 searches the partner information database 5 for partner information and outputs it to the output device 2 (step A5). This makes it possible to browse the partner information in a collective form.

  Next, the operation of the embodiment of FIG. 1 will be described again using a specific example.

  For example, it is assumed that the received e-mail database 4 has an e-mail including an e-mail text “I am a college student!”. Also, in the dictionary 34 for morphological analysis / relationship analysis, the rule that the partner information “x = university student” is extracted from the sentence “x has become a university student” and “profession: university student” (“university student” The rule “Register as“ profession category ”in the information database 5” and the rule “I: Email sender” (register in the partner information database 5 if the actor is “I”) are stored Assume that it is stored.

  First, the control unit 31 of the partner information extraction processing device 3 takes in an e-mail (step A1 in FIG. 2). A morphological analysis process is performed on the electronic mail (step A2). If the e-mail text is "I am a college student!", It is divided into word (morpheme) units such as "I / ha / college student / become /! /!". In the present embodiment, as described above, “Naru” is handled as one word.

  Using the result of this morphological analysis, a relational analysis process is performed (step A3). Based on the morphological analysis result of “I / ha / university student / ni / now /!”, First, the rule that “x = university student” is extracted from the sentence “x became a college student” Thus, the information “I = University student” is extracted. Further, the rule “I: mail sender” indicates that the information “I = university student” is the target information registered in the partner information database 5. Furthermore, it can be understood that the “college student” is registered as the “profession category” in the partner information database 5 by the rule of “occupation: university student”.

  As a result, the partner information that the mail sender is a college student (profession category) is registered in the partner information database 5 with a time stamp (step A4).

  After the above processing, by operating the input device 1 and accessing the partner information database 5, it is possible to browse through the output device 2 information that the current mail sender is a college student (step). A5).

  As described above, according to the present embodiment, the partner information can be automatically extracted from the received electronic mail, and the partner information can be stored in the partner information database 5. There is an effect that it is possible to easily view the partner information in a collective form without performing a search, and the user does not need to manually extract the partner information from the received e-mail. Can be reduced, and the convenience of the user is improved.

  Further, in the present embodiment, a relational analysis process is added to the morphological analysis process, and the morphological analysis / relationship analysis dictionary 34 includes the morpheme analysis process to the extent necessary for the relational analysis process. By registering only the minimum words and rules necessary for relation analysis processing, it is possible to easily mount a natural language processing device even in a small device such as a portable terminal with limited hardware resources. The natural language processing speed can be improved.

  Further, in the present embodiment, by registering the partner information in the partner information database 5 in time series using the time stamp of the e-mail, the partner information can be browsed in time series. Can play.

  Next, with reference to FIG. 3 and FIG. 4, a second embodiment of the mobile terminal equipped with the natural language processing device will be described.

  FIG. 3 is a block diagram showing a second embodiment of a portable terminal equipped with the natural language processing apparatus of the present invention.

  The second embodiment shown in FIG. 3 differs from the first embodiment only in that it has a partner information history processing device 50 instead of the partner information database 5 in the first embodiment shown in FIG. is there. Therefore, in FIG. 3, the same reference numerals or symbols are assigned to the components corresponding to those shown in FIG. 1, and the description thereof is omitted.

  In FIG. 3, the partner information history processing device 50 connected to the control unit 31 has a function of managing and storing partner information as a history, and includes a partner information history management unit 51 and a partner information database 52. Yes.

  The partner information history management unit 51 manages the history of the partner information, manages only the latest information of information belonging to the same category, and registers only the latest information in the partner information database 52. The partner information database 52 is a database in which partner information is registered.

  Next, the operation of the second embodiment shown in FIG. 3 will be described with reference to FIG.

  FIG. 4 is a flowchart for explaining the operation of the second embodiment of FIG. The operation shown in FIG. 4 is different from the operation of the first embodiment only in that the operation of the partner information history management unit 51 is added to the operation of the first embodiment shown in FIG. . Therefore, in FIG. 4, the same reference numerals or symbols are assigned to the components corresponding to those shown in FIG. 2, and the detailed description thereof is omitted.

  In FIG. 4, triggered by reception of the email, the control unit 31 of the partner information extraction processing device 3 takes in the email received from the received email database 4 (step A <b> 1 in FIG. 4), and the morphological analysis unit 32. Then, the electronic mail text is received from the control unit 31 and morpheme analysis processing is performed using the morphological analysis / relationship analysis dictionary 34 (step A2), and the relationship analysis unit 33 displays the result of the morpheme analysis from the control unit 31. Receiving, the morphological analysis / relationship analysis dictionary 34 is used to perform the relationship analysis process, and the partner information extraction process is performed (step A3). Then, the result of the relationship analysis performed by the relationship analysis unit 33 is passed to the control unit 31. The operation so far is the same as the operation from step A1 to step A3 shown in FIG.

  Next, the control unit 31 passes the partner information as a result of the relationship analysis to the partner information history management unit 51, and the partner information history management unit 51 performs history processing of the partner information (step B1). The history processing of the partner information means that when the partner information (hereinafter referred to as the latest partner information) as a result of the relationship analysis is registered in the partner information database 52, existing information belonging to the same category as the latest partner information is It is confirmed whether it is registered in the partner information database 52. If the existing information is registered, the latest partner information is registered instead of the existing information (step A4). As a result, only the latest partner information can be viewed as partner information instead of the past partner information. If the existing information has not been registered, the latest partner information is added and registered in the partner information database 52 as in the operation of the first embodiment (step A4).

  After the above processing, by operating the input device 1 and accessing the partner information database 52, the control unit 31 searches the partner information database 52 for partner information and outputs it to the output device 2 (step A5). As a result, the partner information belonging to the same category can always browse only the latest partner information.

  Next, the operation of the second embodiment will be described again using a specific example.

  For example, it is assumed that existing partner information having “high school student” as “profession category” is registered in the partner information database 52.

  As in the specific example of the first embodiment, it is assumed that the received e-mail database 4 includes an e-mail including an e-mail text “I am a college student!”. Also, in the dictionary 34 for morphological analysis / relationship analysis, the rule that the partner information “x = university student” is extracted from the sentence “x has become a university student” and “profession: university student” (“university student” The rule “Register as“ profession category ”in the information database 5” and the rule “I: Email sender” (register in the partner information database 5 if the actor is “I”) are stored Assume that it is stored.

  The information “I = university student” is extracted by the operations of step A1 to step A3 in FIG. In addition, the rule “I: mail sender” indicates that the information “I = university student” is the target information registered in the partner information database 52. Furthermore, it can be understood that the “college student” is registered as the “profession category” in the partner information database 52 by the rule of “profession: university student”.

  As a result, the partner information history management unit 51 in the partner information history processing apparatus 50 performs confirmation processing to check whether existing information is registered for the “profession category” in the partner information database 52. At this time, it becomes clear that “high school student” is registered as the “profession category”, but based on the time stamp, it is determined that it is the past partner information (step B1 in FIG. 4).

  As a result, in the “profession category” of the partner information database 52, the “high school student” that is the past partner information is deleted, and is registered again as the “university student” that is the latest partner information (step in FIG. 4). A4).

  After the above processing, by operating the input device 1 and accessing the partner information database 52, the current mail sender browses only the information that he is not a high school student but a college student through the output device 2. (Step A5 in FIG. 4).

  As described above, according to the second embodiment of the present invention, since the partner information history processing device 50 performs the partner information history process, when browsing the partner information, only the latest partner information is browsed. The effect that it becomes possible can be produced.

It is a block diagram which shows one Embodiment of the portable terminal carrying the natural language processing apparatus of this invention. It is a flowchart explaining operation | movement of embodiment of FIG. It is a block diagram which shows 2nd Embodiment of the portable terminal carrying the natural language processing apparatus of this invention. It is a flowchart explaining the operation | movement of 2nd Embodiment of FIG.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Input device 2 Output device 3 Counterpart information extraction processing device 31 Control part 32 Morphological analysis part 33 Relational analysis part 34 Morphological analysis / relational analysis dictionary 35 Storage part 4 Received e-mail database 5 Counterpart information database 50 Counterpart information history processing apparatus 51 Partner information history management unit 52 Partner information database 9 Mobile terminal

Claims (12)

A natural language processing device for extracting partner information including attributes and status of a partner with whom an email is exchanged from a received email and registering it in the partner information database,
The natural language processing device includes a morpheme analyzer that performs processing to divide a sentence included in a received email into words.
A relation analysis unit that performs processing to extract partner information based on the analysis result of the morphological analysis unit;
A morpheme that is stored as the partner information in the partner information database and stores the minimum necessary words and relation analysis rules and is referred to when the morpheme analysis unit and the relationship analysis unit perform analysis / extraction processing. A dictionary for analysis and relation analysis,
Control for causing the morpheme analysis unit to analyze the email body of the received email, control for causing the relationship analysis unit to perform relationship analysis based on the analysis result of the morpheme analysis unit, and execution of the partner information extraction process, and the relationship analysis A control unit for performing the control of adding the e-mail address and time stamp information of the e-mail header information of the received e-mail to the result of the extraction process of the unit and registering the information in the counterpart information database;
A natural language processing apparatus comprising:   The natural language processing apparatus further includes a partner information history management unit that manages partner information as a history, and when the control unit obtains a result of the extraction process of the relation analysis unit, the partner information history management unit To check if the partner information of the same category as the result of the extraction process has already been registered in the partner information database, and if the partner information of the same category has already been registered, the existing partner information is The natural language processing apparatus according to claim 1, wherein control is performed so that the result is registered in place of the result.   When the natural language processing device obtains the browsing request information of the partner information from the input device, it searches the partner information database, extracts the partner information that matches the browsing request information, and outputs it to the output device. The natural language processing apparatus according to claim 1, wherein the natural language processing apparatus is characterized. In a natural language processing apparatus that extracts partner information consisting of attributes and status of a partner with whom an email is exchanged from a received email and registers it in the partner information database.
The natural language processing device includes a morpheme analyzer that performs processing to divide a sentence included in a received email into words.
A relation analysis unit that performs processing to extract partner information based on the analysis result of the morphological analysis unit;
A morpheme that is stored as the partner information in the partner information database and stores the minimum necessary words and relation analysis rules and is referred to when the morpheme analysis unit and the relationship analysis unit perform analysis / extraction processing. And a dictionary for analysis and relation analysis,
The natural language processing device causes the morphological analysis unit to analyze the email body of the received e-mail, and causes the relationship analysis unit to analyze the relationship from the analysis result of the morpheme analysis unit, and executes partner information extraction processing And a step of adding a mail address and time stamp information of the e-mail header information of the received e-mail to the result of the extraction process of the relation analysis unit and registering the result in the partner information database. A natural language processing method characterized.   The natural language processing device further includes a partner information history management unit that manages partner information as a history, and the partner information history is obtained when the natural language processing device obtains the extraction processing result of the relation analysis unit. The management unit confirms whether the partner information of the same category as the result of the extraction process has already been registered in the partner information database. If the partner information of the same category has already been registered, the existing partner information is extracted. The natural language processing method according to claim 4, wherein the registration is performed by replacing the result of the processing.   When the natural language processing device obtains the browsing request information of the partner information from the input device, the natural language processing device performs a step of searching the partner information database to extract partner information that matches the browsing request information and outputting it to the output device The natural language processing method according to claim 4, wherein the natural language processing method is any one of claims 4 and 5. A morpheme analyzer that performs processing to divide a sentence included in the received e-mail into words;
A relational analysis unit that performs processing to extract partner information consisting of attributes and status of a partner who is exchanging e-mails based on the analysis result of the morphological analysis unit;
Morphological analysis that is stored in the partner information database as partner information and stores the minimum necessary words and relation analysis rules, and is referred to when the morpheme analysis unit and the relationship analysis unit perform analysis / extraction processing A relational analysis dictionary, and a computer of a natural language processing device that extracts partner information from received e-mail and registers it in the partner information database;
A process for causing the morphological analysis unit to analyze the body of the email of the received e-mail, a process for causing the relational analysis unit to perform a relational analysis based on an analysis result of the morpheme analysis part, and a process for executing the partner information extraction process And a process of registering in the partner information database with the mail address and time stamp information of the e-mail header information of the received e-mail added to the result of the part extraction process.   The natural language processing apparatus further includes a partner information history management unit that manages partner information as a history, and the program stores the partner information history when the computer obtains a result of extraction processing from the relation analysis unit. The management unit confirms whether the partner information of the same category as the result of the extraction process has already been registered in the partner information database. If the partner information of the same category has already been registered, the existing partner information is extracted. The program according to claim 7, wherein a program for performing a registration by replacing with a result of the processing is executed.   The program executes a process of searching the partner information database, extracting partner information that matches the browse request information, and outputting the partner information database to the output device when the partner information browse request information is obtained from the input device. The program according to any one of claims 7 and 8. A mobile terminal equipped with a natural language processing device that extracts partner information including attributes and status of the partner with whom the email is exchanged from the received email and registers it in the partner information database,
The natural language processing device includes a morpheme analyzer that performs processing to divide a sentence included in a received email into words.
A relation analysis unit that performs processing to extract partner information based on the analysis result of the morphological analysis unit;
A morpheme that is stored as the partner information in the partner information database and stores the minimum necessary words and relation analysis rules and is referred to when the morpheme analysis unit and the relationship analysis unit perform analysis / extraction processing. A dictionary for analysis and relation analysis,
Control for causing the morpheme analysis unit to analyze the email body of the received email, control for causing the relationship analysis unit to perform relationship analysis based on the analysis result of the morpheme analysis unit, and execution of the partner information extraction process, and the relationship analysis A control unit for performing the control of adding the e-mail address and time stamp information of the e-mail header information of the received e-mail to the result of the extraction process of the unit and registering the information in the counterpart information database;
A portable terminal equipped with a natural language processing device.   The natural language processing apparatus further includes a partner information history management unit that manages partner information as a history, and when the control unit obtains a result of the extraction process of the relation analysis unit, the partner information history management unit To check if the partner information of the same category as the result of the extraction process has already been registered in the partner information database, and if the partner information of the same category has already been registered, the existing partner information is The mobile terminal equipped with the natural language processing device according to claim 10, wherein control is performed to register the result by replacing the result. When the natural language processing device obtains the browsing request information of the partner information from the input device, it searches the partner information database, extracts the partner information that matches the browsing request information, and outputs it to the output device. A portable terminal equipped with the natural language processing device according to any one of claims 10 and 11.

Images