JP2010061508A - History information retrieval apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To extract/store characteristic causality patterns from a plurality of text data groups having date information, such as news articles, to analyze text data specified by a user and to retrieve other deeply related articles based on the stored causality patterns. <P>SOLUTION: A history information retrieval apparatus includes: a term information storage means; an article concept information storage means; an article information storage means; a history information storage means; a conceptualization means for segmenting a keyword from the text data of article information or text data received from a user and conceptualizing the segmented keyword; an article concept information preparation means for conceptualizing the text data of all article information stored in the article information storage means; a history information retrieval means for conceptualizing the text data received from the user, collating the conceptualized text data with the patterns of a series of events generated in the past and retrieving other articles with historically deep relation; and an output means for presenting a retrieval result obtained by the history information retrieval means to the user. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a history information search apparatus, and more particularly, extracts and stores characteristic causal patterns from a plurality of text data groups having date and time information such as news articles, and further, text data designated by a user. And a technique for searching other articles that are closely related based on the accumulated causal pattern.

  2. Description of the Related Art Conventionally, as described in Patent Document 1 below, an article summarizing device that generates summaries of a plurality of news articles has been proposed.

As prior art documents related to the invention of the present application, there are the following.
JP-A-2005-250648

According to the invention described in Patent Document 1 described above, an article that matches a user search condition and a plurality of articles that are closely related to the article are extracted, and a summary of the extracted related articles is generated and presented. can do.
However, the related article in the invention described in the above-mentioned Patent Document 1 is an article editor that associates a connection of a subsequent article with a new article in advance. There was a problem that it was not possible to find articles that are related to historical causal relationships that were not intended.
The present invention has been made to solve the problems of the prior art, and an object of the present invention is to create a characteristic causal pattern from a plurality of text data groups having date and time information such as news articles. To provide a history information search device that can extract and store, analyze text data specified by a user, and search other articles that are closely related based on the stored causal pattern is there.
The above and other objects and novel features of the present invention will become apparent from the description of this specification and the accompanying drawings.

Of the inventions disclosed in this application, the outline of typical ones will be briefly described as follows.
In order to solve the above-described problems, the present invention is a historical information search apparatus, and is term information that is a correspondence table for replacing specific terms appearing in article text with words or short sentences representing abstract concepts. Term information storage means for storing information, article concept information storage means for storing article concept information that expresses the content of individual article information in an abstract concept, and article information that combines date information and text information that represents the contents Article information storage means for storing information, history information storage means for storing history information consisting of a plurality of article information representing a series of events that occurred in the past, and names representing the entire series of events, and text data of article information Alternatively, a conceptualization unit that cuts out keywords from text data received from a user and conceptualizes the text data of all article information stored in the article information storage unit. Historical information that searches for other historically relevant articles by conceptualizing the text data received from the user and collating with a series of events that occurred in the past It is characterized by comprising search means, input means for receiving text data in response to an input from the user, and output means for presenting the search result obtained by the history information search means to the user.

The effects obtained by the representative ones of the inventions disclosed in the present application will be briefly described as follows.
According to the history information search apparatus of the present invention, it is possible to search for history information closely related to the contents of text data input by the user. You can also predict what will happen in the future.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.
In all the drawings for explaining the embodiments, parts having the same functions are given the same reference numerals, and repeated explanation thereof is omitted.
[Function block]
FIG. 1 is a functional block diagram of a history information retrieval apparatus according to an embodiment of the present invention.
As shown in FIG. 1, the history information search apparatus of this embodiment includes term information storage means 101, article concept information storage means 102, article information storage means 103, history information storage means 104, conceptualization means 105, and article concept information creation. Means 106, history information search means 107, input means 108, and output means 109 are provided.
The term information storage unit 101 stores a plurality of term information. The term information is expressed in a correspondence table for replacing specific terms appearing in the article text with words or short sentences representing abstract concepts.
The article concept information storage unit 102 stores a plurality of article concept information. The article concept information is an abstract concept representing the contents of individual article information stored in the article information storage means 103.
The article information storage unit 103 stores a plurality of article information. The article information is a pair of date information and text information representing the contents, such as news and newspapers.
The history information storage unit 104 stores a plurality of history information. The history information is composed of a plurality of pieces of article information representing a series of events that have occurred in the past and a name representing the entire series of events.

The conceptualization means 105 receives the text data of the article information taken out from the article information storage means 103 by the article concept information creation means 106 or the text data received from the input means 108 by the history information search means 107. A keyword is cut out from the received text data and collated with the term information stored in the term information accumulating means 101 to conceptualize the text data.
The article concept information creating unit 106 conceptualizes the text data of the article information by the conceptualizing unit 105 for all article information stored in the article information storing unit 103. Further, the result is stored in the article concept information storage means 102.
The history information search means 107 conceptualizes the text data received from the input means 108 by the conceptualization means 105. Based on the result, similar history information and article information included in the history information are searched from the article concept information storage means 102, the article information storage means 103, and the history information storage means 104, and output by the output means 109. Present the result.
The input means 108 receives text data that is input from the user and passes it to the history information search means 107. The output unit 109 receives the result searched by the history information search unit 107 and presents it to the user.

FIG. 2 shows an example of a hardware configuration in which the history information retrieval apparatus of this embodiment is realized using a CPU.
The history information search apparatus according to the present embodiment includes a display 201, a CPU 203, a memory 205, a keyboard / mouse 207, a hard disk 209, a CD-ROM drive 211, and a communication circuit 215.
The hard disk 209 stores a conceptualization program 2091, an article concept information creation program 2092, a history information search program 2093, and data necessary for the information search process of the present embodiment for performing the information search process of the present embodiment. Term information database 2094, article information database 2095, article concept information database 2096, history information database 2097, and operating system 2098 are recorded. These are installed by reading data recorded on the CD-ROM 213 via the CD-ROM drive 211. The installation may be performed using data downloaded from the Internet 217 or the like using the communication circuit 215.

Hereinafter, the data structure of various information used in the information search process of the present embodiment will be described.
(3-1) Term Information FIG. 3 shows a data structure of term information and an example thereof. The term information has two columns of a term (301) and a concept name (302).
The term (301) describes a specific term that appears in the text of an article.
The concept name (302) describes a word representing a concept of a specific term stored in the term (301) or a short sentence.
For example, FIG. 3 shows four terms information. Of these, the first line shows that the specific term “stock” expresses the concept of “value of financial assets”.
(3-2) Article Information FIG. 4 shows the data structure of article information and an example thereof. The article information has three columns of article ID (401), date and time (402), and text (403).
The article ID (401) describes an identifier for uniquely identifying article information in the historical information search apparatus of the present embodiment.
Date and time (402) describes the date and time when the event of the article occurred. In the example of FIG. 4, since news articles are used, information up to the year, month, and day is described, but it may be described more precisely to the accuracy of hours, minutes, and seconds.
The body (403) describes text data representing the content of the article.
For example, FIG. 4 illustrates three pieces of article information. Of these, the article information with the article ID “K000001” indicates that the date and time is “October 15, 2007”, and the body text is “the stock price of OO company suddenly drops”.

(3-3) History Information FIG. 5 shows a data structure of history information and an example thereof. The history information has four columns of history ID (501), history name (502), article order (503), and article ID (504).
The history ID (501) describes an identifier for uniquely identifying history information in the history information search apparatus of the present embodiment.
The history name (502) describes the name of the history information.
The article order (503) represents a time series order of a plurality of article information included in the history information. The article order is expressed as a positive integer starting from 1.
Article ID (504) describes the IDs of a plurality of article information included in the history information.
For example, FIG. 5 illustrates one piece of historical information having the name “Heisei Depression” and another piece of historical information having the name “Black Monday”. That is, in the first three lines, the history ID is “R000001”, the history name is “Heisei Depression”, and the history information includes the article ID “K000001”, the article ID “K000002” This indicates that an article with ID “K000003” is included in this order. (3-4) Article Concept Information FIG. 6 shows the data structure of article concept information and an example thereof. The article concept information has three columns of article ID (601), subject (602), and phenomenon (603).
The article ID (601) is the article ID (401) of the article information described in the above (3-2).
The subject (602) describes the concept indicating the subject in the content of the article information indicated by the article ID (601).
The phenomenon (603) describes the concept indicating the phenomenon in the content of the article information indicated by the article ID (601).
For example, FIG. 6 illustrates three pieces of article concept information. Among these, in the article concept information represented by the first row, the content of the article information whose article ID is “K000001” represents that “the value of the financial asset” (subject) has been “decreased” (phenomenon). ing.

The processing of each program shown in FIG. 2 will be described below using the flowcharts of FIGS.
(4-1) Conceptualization Program FIG. 7 is a flowchart showing the processing procedure of the conceptualization program 2091 shown in FIG. The processing procedure of the conceptualization program 2091 shown in FIG. 2 will be described below using the flowchart of FIG.
Step 701: The input text is analyzed to obtain a portion corresponding to a main phrase and a predicate of a sentence. Examples of this analysis method include a method using morphological analysis or syntax analysis processing described in JP-A-2005-149359. Another example is a method using a dependency analysis program “Cabocha” developed at Nara Institute of Science and Technology.
Step 702: Search the term information database 2094 for term information including a keyword included in the main phrase acquired in step 701.
Step 703: Search the term information database 2094 for term information including the keyword included in the predicate acquired in step 701. As described above, the concept name of the term information obtained in step 702 is mainly used, and the concept name of the term information obtained in step 703 is returned as an output to the upper program as a phenomenon.

(4-2) Article Concept Information Creation Program FIG. 8 is a flowchart showing a processing procedure of the article concept information creation program 2092 shown in FIG. The processing procedure of the article concept information creation program 2092 shown in FIG. 2 will be described below using the flowchart of FIG.
Step 801: The total number of records of article information stored in the article information database 2095 is acquired.
Step 802: The processing from Step 803 to Step 805 is repeated for the total number of records acquired in Step 801. Hereinafter, the repetition counter is n and the initial value is 1.
Step 803: The data on the nth row is acquired from the article information stored in the article information database 2095.
Step 804: The body text of the article information acquired in Step 803 is input, and the conceptualization program 2091 described in (4-1) is called. From the conceptualization program 2091, conceptual information (subject / phenomenon) is obtained as an output.
Step 805: One record of article concept information is created based on the article ID of the article information acquired in step 803 and the concept information (subject / phenomenon) acquired in step 804, and stored in the article concept information database 2096.
Step 806: It is an end of repetition corresponding to Step 802. The repeat counter n is incremented.

(4-3) History Information Search Program FIG. 9 is a flowchart showing the processing procedure of the history information search program 2093 shown in FIG. The processing procedure of the history information search program 2093 shown in FIG. 2 will be described below using the flowchart of FIG.
Step 901: The text data from the user obtained from the input means 108 is input, and the conceptualization program 2091 described in (4-1) is called. From the conceptualization program 2091, conceptual information (subject / phenomenon) is obtained as an output.
Step 902: Article concept information having concept information matching the concept information (subject / phenomenon) acquired in step 901 is searched from the article concept information database 2096.
Step 903: The processing from step 904 to step 911 is repeated by the number of search results in step 902. Hereinafter, the repetition counter is n1, and the initial value is 1.
Step 904: The article concept information corresponding to the n1th line is acquired from the article concept information database 2096 among the search results of Step 902.
Step 905: The history information including the article ID of the article concept information acquired in Step 904 is searched from the history information database 2097.
Step 906: The process of step 907 is repeated by the number of search results of step 905. Hereinafter, the repetition counter is n2, and the initial value is 1.

Step 907: History information corresponding to the n2th line is acquired from the history information database 2097 among the search results of Step 905. Further, all the article information included in the history information is acquired from the article information database 2095.
Step 908: Article concept information corresponding to all articles acquired in step 907 is acquired from the article concept information database 2096.
Step 909: A series of other pieces of article information matching the sequence of all article concept information acquired in step 908 is acquired from the article concept information database 2096 and the article information database 2095.
Step 910: It is the end of the iteration corresponding to step 906. The repeat counter n2 is incremented by one.
Step 911: It is a repetition end corresponding to Step 903. The repeat counter n1 is incremented by one.
Step 912: History information acquired in Step 907, Step 908, and Step 909, all the article information included in the history information, and other series of article information that matches the article concept information corresponding to the history information. The result is presented to the user by the output means 109.

In FIG. 10, an example of the search result of the historical information search device of a present Example is shown.
When the text “XX bankruptcy” is input by the user, the conceptualizing means 105 converts it into conceptual information (subject / phenomena) “company / bankruptcy”.
Concept information that matches this is acquired from the article concept information database 2096, and from the history information database 2097, the history information including the acquired article concept information is included in the history information “Heisei Depression” and the history information. A series of article information is retrieved and presented to the user. (1001 in FIG. 10)
Further, the article concept information database 2096 and the article information database 2095 are searched, and the article concept information corresponding to the previous period of the history information is arranged (“value / decrease in management assets” → “company / bankruptcy” → “management”). Search for other series of article information that matches "index / increase"). This makes it possible to discover the history of the same pattern that occurred in the past. (1002 in FIG. 10)
You can also discover the history of the same pattern that will occur in the future. In this case, there is no article that matches the article concept information “management index / increase” yet, but it can be predicted that an event that matches the article concept information may occur in the future. (1003 in FIG. 10)
In the above-described embodiment, the CPU 203 is used to realize the function shown in FIG. 1, and this is realized by software. However, some or all of them may be realized by hardware such as a logic circuit. In addition, you may make it make an operating system (OS) process a part of program.
As mentioned above, the invention made by the present inventor has been specifically described based on the above embodiments. However, the present invention is not limited to the above embodiments, and various modifications can be made without departing from the scope of the invention. Of course.

It is a figure which shows the functional block diagram of the historical information search device of the Example of this invention. It is a figure which shows an example of the hardware constitutions of the historical information search device of the Example of this invention. It is a figure which shows the data structure and term information of term information of the Example of this invention. It is a figure which shows the data structure and article information of the article information of the Example of this invention. It is a figure which shows the data structure and historical information of the historical information of the Example of this invention. It is a figure which shows the data structure and article concept information of the article concept information of the Example of this invention. It is a flowchart which shows the process sequence of the conceptualization program shown in FIG. It is a flowchart which shows the process sequence of the article concept information creation program shown in FIG. It is a flowchart which shows the process sequence of the flowchart of the historical information search program shown in FIG. It is a figure which shows an example of the search result of the historical information search device of the Example of this invention.

Explanation of symbols

101 Term information storage means 102 Article concept information storage means 103 Article information storage means 104 History information storage means 105 Conceptualization means 106 Article concept information creation means 107 History information search means 108 Input means 109 Output means 201 Display 203 CPU
205 Memory 207 Keyboard / Mouse 209 Hard Disk 211 CD-ROM Drive 213 CD-ROM
215 Communication circuit 301 Term 302 Concept name 401, 504, 601 Article ID
402 Date 403 Full Text 501 History ID
502 History name 503 Article order 602 Subject 603 Phenomenon

Claims (1)

Term information storage means for storing term information, which is a correspondence table for replacing specific terms appearing in the article body with words or short sentences representing abstract concepts,
Article concept information storage means for storing article concept information expressing the content of individual article information in an abstract concept,
Article information storage means for storing article information in which date / time information and text information representing contents are paired;
Historical information storage means for storing a plurality of article information representing a series of events that have occurred in the past, and historical information consisting of names representing the entire series of events,
A conceptualization means for extracting and conceptualizing keywords from text data of article information or text data received from a user;
Article concept information creating means for conceptualizing text data of all article information stored in the article information storage means;
Historical information retrieval means that conceptualizes text data received from users, matches a series of events that occurred in the past, and searches other historically related articles,
An input means for receiving text data in response to an input from the user;
A history information search apparatus comprising: output means for presenting a search result obtained by the history information search means to a user.

Images