JP4034374B2 - Information search system and information search method - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention searches electronic text information such as e-mails, WWW homepages, network news articles, and chat logs that are conversation logs in a network conversation system, and matches the user's request from these. Search system and information search method for selecting a user, and in particular, a search with a high reproducibility is realized by generating a search condition from arbitrarily selected text information and text information having a predetermined relationship with the text information. The present invention relates to an information search system and an information search method.
[0002]
[Prior art]
In recent years, in addition to the widespread use of word processors and personal computers, the spread of electronic mail and WWW (World Wide Web) via computer networks such as the Internet and personal computer communication has been rapidly progressing. As a result, the amount of computerized information that can be held and accessed (obtained) by individuals has become enormous, and the effort to search for necessary information has increased.
[0003]
From such a background, there is an increasing demand for a so-called information retrieval system that selects an electronic text information possessed by an individual or publicly available to meet a user's request / interest.
[0004]
However, the information retrieval system that has been realized in the past mainly adopts a method of searching for a conforming document based on the presence or absence of a word (or its synonym) input by a user, and a single or plural designations are adopted. It was not suitable for searching electronic text information similar to documents and topics (electronic text information continuous in time).
[0005]
[Problems to be solved by the invention]
As described above, the conventional information retrieval system is not suitable for retrieving electronic text information similar to one or a plurality of designated documents or topics (electronic text information continuous in time). there were.
[0006]
The present invention has been made in view of such circumstances, and uses electronic text information in the vicinity of one or more electronic text information specified by the user for retrieval, thereby resembling the electronic text information. It is an object of the present invention to provide an information search system and an information search method capable of improving the recall of electronic text information search and enabling search in accordance with the topic. Furthermore, the present invention provides an information retrieval system for preventing a decrease in precision by controlling the development of synonyms executed at the time of retrieval based on designated electronic text information and nearby electronic text information, and The object is to provide an information retrieval method.
[0007]
[Means for Solving the Problems]
In order to achieve the above-described object, the present invention provides a text information selection means for selecting one or more text information from a plurality of text information in an information search system for searching various text information, and the text A relational text information collecting means for collecting text information existing within a predetermined range from the text information selected by the information selecting means; a text information selected by the text information selecting means; and a relational text information collecting means. Search condition generation means for generating a search condition from the text information obtained, and text information search means for searching for text information that matches the search condition generated by the search condition generation means. .
[0008]
DETAILED DESCRIPTION OF THE INVENTION
An embodiment of the present invention will be described below with reference to the drawings.
FIG. 1 shows a schematic configuration of an information device on which an information search system according to an embodiment of the present invention operates. As shown in FIG. 1, the information apparatus on which the information retrieval system of this embodiment operates includes a central processing unit 1 that controls the entire apparatus and performs various processes, a semiconductor memory that stores programs and data, a magnetic disk, an optical disk, and the like. A storage device 2, a communication device 3 for accessing information sources such as the Internet and a personal computer communication service system, an input device 4 such as a keyboard and a mouse for receiving instructions from a user, and a monitor for transmitting information to the user Output device 5. That is, the information retrieval system of this embodiment operates on a very general personal computer or an information device having a function equivalent to or better than that.
[0009]
FIG. 2 shows a functional configuration of the information search system of this embodiment. As shown in FIG. 2, the information search system of this embodiment includes a text information selection unit 11 for selecting text information as a seed for generating a search command, and a user interface unit 12 for control by the user. A text information storage unit 13 for storing text information held by the user, a news article access unit 14 for accessing network news, a WWW access unit 15 for accessing a WWW homepage, and a predetermined relationship. A neighboring text information selection unit 16 for selecting text information (neighboring text information), a parameter storage unit 17 for storing parameters used when selecting the neighboring text information, and a search command generation for generating a search command from the text information Part 18 and controls the development of synonyms from text information A synonym expansion control unit 19, a synonym dictionary storage unit 20 for storing a synonym dictionary, a text information search unit 21 that searches text information suitable for a search command from a set of text information, and a search result as a user The search result display part 22 presented in FIG.
[0010]
FIG. 3 shows a processing flow of the information search system of this embodiment.
As shown in FIG. 3, the processing of this information retrieval system selects electronic text information in the vicinity of text information designated by the user (step A1), and generates a search command from the text information obtained in step A1. (Step A2), synonym expansion of words of the search command based on the text information obtained in Step A1 (Step A3), and retrieval of text information by the words obtained by this expansion (Step A4). Hereinafter, each processing will be described in detail.
[0011]
FIG. 4 shows details of the processing in step A1 of FIG. In this information retrieval system, for each selected text information (which may be one or more),
(1) If the designated text information is a WWW homepage, a homepage in the vicinity of the homepage is searched (step B3).
[0012]
(2) If the designated text information is a network news article, search for a news article near the news article (step B5).
(3) If the designated text information is an electronic mail, an electronic mail in the vicinity of the electronic mail is searched (step B7).
[0013]
(4) If the designated text information is a chat log, the log range in the vicinity of the designated position of the chat log is determined. (Step B9).
Here, “neighborhood” varies depending on the definition of “distance” that can be placed in the text information. First, the “distance” in each text information is defined here.
[0014]
(1) The distance in the case of a WWW home page is defined as the minimum number of links required to reach from a home page to another page.
[0015]
(2) In the case of network news articles, when articles in a thread (following article chain) are arranged in chronological order, the distance between two articles in the same thread is the difference in the order given to each article. Define (the distance between articles in different threads is infinite).
[0016]
(3) In the case of e-mail, the same as in the case of network news articles.
(4) In the case of a chat log, the distance between two positions is defined as either the number of characters / number of sentences / number of paragraphs included between the two positions or the time difference between the two positions. Can be selected.
[0017]
Whether or not the two pieces of text information are “near” is determined by whether or not the distance between the two pieces of text information is within a certain threshold. In this information retrieval system, this threshold value can be stored in the parameter storage unit 17, and the value can be changed at the user's discretion.
[0018]
FIG. 5 shows details of the process in step A2 of FIG. In this information search system, it is assumed that the search command is a set of words that can be determined as to whether the document is a relevant document.
[0019]
First, the information retrieval system cuts out a word from a user's designation or each nearby text information by a technique such as morphological analysis (step C1). Next, the words thus obtained are rearranged in descending order of frequency (step C2), and words having a frequency equal to or lower than the parameter (threshold value) stored in advance in the parameter storage unit 17 are deleted (step C3). Then, the word remaining by this procedure is used as a search command. This parameter can be changed by the user.
[0020]
FIG. 6 shows details of the process in step A3 of FIG. Here, a specific example will be described first.
The synonym dictionary format is such that a list of pairs of synonyms and concepts to which the synonyms belong is linked to “headwords”. As a form of expression, for example, the form of FIG. 7 can be considered. Here, it is assumed that the contents shown in FIG. 8 are stored in this synonym dictionary. Moreover, the arrow represents the pointer for linking each synonym node. Then, consider the synonym expansion of the word “CD” for such a dictionary.
[0021]
"CD" in the synonym dictionary is
Concept A
Cash dispenser cash dispenser cash dispenser
Concept B
Compact disc compact disc
Concept C
Transferable deposits concept D
Conference on Disarmament
Suppose that “CD” is expanded into a total of 11 words classified into four concepts.
[0022]
At this time, which concept synonym is selected depends on which concept word contains a lot of text information specified by the user and text information in the vicinity thereof. If there are many words of concept B in the text information, “compact disc” or “compact disc” is adopted as a synonym for “CD”.
[0023]
Based on this specific example, details of synonym development will be described with reference to FIG. This information retrieval system first expands a synonym of a word for each word of a search command (step D2), and text information designated by the user for the type of concept obtained by synonym expansion and its vicinity The concept information contains many words of which concept (step D3), and synonyms belonging to the most existing concept are adopted (step D4).
[0024]
These “neighboring” text information groups may constitute one topic in some cases. Therefore, as in this information search system, it is possible to search according to the topic by using the nearby text information, and the word included in the nearby text information is more than the word included in the original electronic text information. As the number increases, the reproduction rate of the search can be improved.
[0025]
Further, by controlling the synonym expansion, it is possible to eliminate a search by a word belonging to an unnecessary concept and avoid a search for unnecessary text information, thereby preventing a decrease in precision.
[0026]
In the information retrieval system of the above-described embodiment, synonym expansion is controlled by text information specified by the user and electronic text information information in the vicinity thereof, but all synonyms included in the synonym dictionary are directly used. Even if it is adopted, it is effective because the recall rate of search can be improved.
[0027]
Moreover, even if it is not necessary to determine text information in the vicinity of electronic text information designated by the user, if the synonym development is controlled by the text information processing designated by the user, it is possible to prevent the precision from being lowered. It is possible and effective.
[0028]
【The invention's effect】
As described above in detail, according to the present invention, electronic text information in the vicinity of text information designated by the user is used as information for creating a search command, so that the search reproducibility is increased and the search is adapted to the topic. Can be made possible.
[0029]
In addition, by controlling synonym development based on text information specified by the user and electronic text information in the vicinity, it is possible to prevent a decrease in the search precision.
[Brief description of the drawings]
FIG. 1 is a diagram showing a schematic configuration of an information device on which an information search system according to an embodiment of the present invention operates.
FIG. 2 is an exemplary diagram showing a functional configuration of the information search system according to the embodiment.
FIG. 3 is an exemplary flowchart showing the flow of processing of the information search system of the embodiment;
4 is a flowchart showing details of processing in step A1 in FIG. 3;
FIG. 5 is a flowchart showing details of processing in step A2 of FIG. 3;
FIG. 6 is a flowchart showing details of processing in step A3 of FIG. 3;
FIG. 7 is a conceptual diagram showing a format of a synonym dictionary according to the embodiment.
FIG. 8 is a diagram illustrating a specific example of the stored contents of the synonym dictionary of the embodiment.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 1 ... Central processing unit, 2 ... Memory | storage device, 3 ... Communication apparatus, 4 ... Input device, 5 ... Display apparatus, 11 ... Text information selection part, 12 ... User interface part, 13 ... Text information storage part, 14 ... News article Access unit, 15 ... WWW access unit, 16 ... Neighboring text information selection unit, 17 ... Parameter storage unit, 18 ... Search command generation unit, 19 ... Synonym expansion control unit, 20 ... Synonym dictionary storage unit, 21 ... Text information Search unit, 22 ... Search result display unit.

Claims (10)

In an information search system that searches various text information,
Text information selection means for selecting one or more text information from a plurality of text information;
Relevant text information collecting means for searching a specified threshold range from text information selected by the text information selecting means and collecting text information existing in the range ;
Search condition generating means for generating a search condition from the text information selected by the text information selecting means and the text information collected by the related text information collecting means;
An information search system comprising: text information search means for searching for text information that matches the search conditions generated by the search condition generation means. When the text information selected by the text information selection unit is hypertext type data, the related text information collection unit displays the text information connected to the text information within the number of links specified as the threshold value. The information search system according to claim 1, wherein the information search system is collected. The related text information collecting means, when the text information selected by the text information selecting means is data having a generation order or generation time, a difference in order or time given when the text information is arranged in time series 2. The information search system according to claim 1, wherein the text information within the value specified as the threshold value is collected from the text information selected by the text information selection means. When the text information selected by the text information selection unit is log data, the related text information collection unit includes the number of characters, the number of sentences, or the paragraph interposed between the text information selected by the text information selection unit. 2. The information retrieval system according to claim 1, wherein text information within a number specified as the threshold value is collected.   Controlling synonym expansion of search terms included in the search condition generated by the search condition generation means using the text information selected by the text information selection means and the text information collected by the related text information collection means The information retrieval system according to claim 1, further comprising synonym expansion control means. In an information search method for a computer, which searches various text information,
The computer is
Accepts input to select one or more text information from multiple text information,
Search the specified threshold range from the selected text information, collect the text information existing in the range ,
Generating a search condition from the selected text information and the collected text information;
Searching for text information that matches the generated search condition. The related text information is collected by the computer when the selected text information is hypertext type data, and the text information connected with the text information within the number of links specified as the threshold is collected. The information retrieval method according to claim 6, wherein the information retrieval method is a thing. The related text information is collected by the computer when the selected text information is data having a generation order or generation time, or when the text information is arranged in time series, the difference in order or time difference is given. The information search method according to claim 6, wherein text information within a value specified as the threshold value is collected with the selected text information. In the collection of the related text information by the computer , when the selected text information is log data, the number of characters, the number of sentences or the number of paragraphs intervening with the selected text information is used as the threshold value. 7. The information search method according to claim 6, wherein text information within a specified number is collected. The computer further controls synonym expansion of a search term included in the generated search condition using the selected text information and the collected text information. Information retrieval method.

Images