JP3135422B2 - Document retrieval apparatus and method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a document search apparatus, and more particularly to a full-text search apparatus for performing a search on all text.

[0002]

2. Description of the Related Art With the spread of a document database and an improvement in computer processing capability, a document retrieval apparatus for retrieving a document including a specified keyword from a large number of document databases has been used.

[0003] In particular, recently, in order to improve the flexibility in searching, a free keyword can be specified by searching the entire text instead of using a keyword attached to a document in advance. The full text search method has been used. This method has a feature that anyone can search because there is no restriction that any keyword must be used.

In a full-text search, a concept represented by a keyword specified by a user (hereinafter, referred to as a search word) may appear as another expression in a document. It is common practice to add a synonym / heterograph as a search word and use it (this is called a search word expansion) to perform a search.

[0005]

By the way, when searching for an object represented by a proper noun, its name may be changed depending on time. For example, when a company name or a country name is changed. Also, when searching for a person, even if their first and last names are not changed, people who hold a position are often referred to by their last name + title or title only, rather than their full names. In such an expression form, the title changes depending on the position at different times.

[0006] The conventional synonym dictionary does not support such names that change with time. In order to deal with this, it is conceivable to register all names used in the past as synonyms, but if you do so, documents that have not been used at a certain time will be searched using that name. However, there is a risk that unnecessary documents may be obtained. In addition, in the case of a person, it is conceivable to use only the last name as a search term, but other people with the same last name will also be searched,
After all, the noise at the time of search increases. In any case, in the case of the conventional document search apparatus, there is a problem that the search accuracy is reduced for an object having a name that changes depending on the time.

An object of the present invention is to eliminate the above-mentioned drawbacks and improve the search accuracy of a document search device.

[0008]

In order to solve the above problems, a document search apparatus according to the present invention includes a document storage means for storing a document together with date information for the document, a history of names of objects, and a record of each name. Name history storage means for storing in association with the use period of the name, input means for inputting a search word,
A search term expansion unit that expands the search word by referring to the name history storage unit based on the search word input from the input unit and the date information of the search target document stored in the document storage unit; Search means for searching the search target document by the search word expanded by the search word expansion means.

According to another aspect of the present invention, in a document search method for searching for a document in a document file that stores the document together with date information for the document by using a search term, a history of an object name is stored. Each name is stored in the name history memory in association with the use period of the name, a search word is input, and the input search word and the date information of the search target document stored in the document file are stored. , The search word is expanded by referring to the name history memory, and the search target document is searched by the expanded search word.

[0010]

The above-mentioned document search device refers to the name history storage means based on the search word input from the input means and the date information of the search target document stored in the document storage means, and searches the search word expansion means. And the search means searches the search target document based on the expanded search word.

According to the above document search method, a history of names of things is stored in a name history memory in which each name is associated with a use period of the name, and a search word is input. The search term is expanded by referring to the name history memory based on the search term and the date information of the search target document stored in the document file, and the search target is expanded by the expanded search term. Search for documents.

[0012]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described below in detail with reference to the drawings.

FIG. 1 is a block diagram showing the configuration of an apparatus according to one embodiment of the present invention. In this embodiment, a case where a search for a personal name is performed is taken as an example.

In FIG. 1, reference numeral 101 denotes a search word input unit for inputting a search word, 102 denotes a search word expansion unit for adding a synonym or the like to the input search word, and 103 denotes a search word including the expanded word. The retained search term storage unit 104 is a synonym dictionary used for search term expansion. 105 is a job history database describing the history of jobs for each person; 106 is a job history holding unit that holds the job history of the person to be searched; 107 is whether or not the document contains a search term A search processing unit, 108 is a document database storing documents, 109 is a document date determination unit that checks the date of the document, 110 is a search result determination unit that checks whether the document obtained as a result of the search is correct, and 111 is a post name. A post determination rule storage unit 112 for storing rules for determining whether the search result is correct is a search result storage unit for storing search results.

FIG. 2 illustrates the present invention more specifically.

In FIG. 1, reference numeral 201 denotes a control memory for storing the control procedure shown in FIG. This may be a ROM or a RAM. Reference numeral 202 denotes a central processing unit that performs processing according to a control procedure stored in the control memory 201. Reference numeral 203 denotes a memory, a search word holding unit 103, a search result holding unit 112, and a post history holding unit 1.
06. There is an area used as the post determination rule holding unit 111. A keyboard 204 is used by a user to input a search word. A disk device 205 has a document database, a post history database, and a synonym dictionary. Reference numeral 206 denotes a display, such as a CRT or a liquid crystal display. This is used to display search results. 207 is a bus.

FIG. 3 is a flowchart showing a processing procedure of the operation in the apparatus shown in FIG. The operation of the embodiment of the present invention will be described with reference to FIG. Here, a case will be described in which only one search word is input. It is also assumed that documents are stored in the document database 108 in chronological order.

First, in step S301, the position history database 105 is searched for the search word input from the search word input unit 101, and the position history of the person of the search word obtained as a result of the search is stored in the position history holding unit 106. Hold. In the post history database 105, for example, a period and a post (or title) are described as shown in FIG. Among them, "previous prime minister" and "former prime minister" are not titles but are used as titles, so such a prime minister is also described.

Next, in step S302, the retrieval word holding unit 10
3 is initialized. The search word holding unit 103 holds the presence / absence, period, and search word of the post history. If there is a post history, the date up to the day before the start date of the oldest post is described as the initial value of the period, and the first and last name without the post is used as the initial value of the search word. For example, if the start date of the oldest post of "Noboru Takeshita" is 05/21/1958, as shown in (1) of FIG.
05/20], and the search term is “Takeshita Noboru”. If there is no post history, enter the search word as it is. For example, when there is no post history of "Taro Nippon", the post history is set to "none" and the search word is "Taro Nippon" as shown in (2) of FIG.

In step S303, one document is extracted from the document database 108. Next, step S304
In the search word holding unit 103, it is checked whether or not the post history is "present". If there is no post history, a normal search may be performed, and the process proceeds to step S311.
A search process is performed to check whether a search term is included in the document.

If there is a post history in step S304, the process proceeds to step S305, and the document date determination unit 109 determines whether the date of the document of interest is within the period held in the search word holding unit 103. Find out. If it is within the period, the process moves to step S307, and a search process is performed using the search word currently held in the search word holding unit 103. If it is not within the period, the job title needs to be changed, so the process moves to step S306, where the job title corresponding to the date of the document is extracted from the job history holding unit 106, and the search term is further expanded. When the search word is expanded, the synonym dictionary 104 is used to look up synonyms of the job title, and each is searched for <surname +
Expands to <title> and <title>. The expanded result and the original search word <first name and last name> are held in the search word holding unit 103 together with the period. Then, a search process is performed in step S307.
The search processing in step S307 is the same as the search processing in step S311. It is checked whether the search term held in the search term holding unit 103 exists in the document. This method may be in accordance with a generally known character string collation algorithm, or an index may be prepared in advance and searched using the index.

In step S308, if there is a search word in the document, it is checked whether or not there is a word having only the title. If not, the process moves to step S310. If there is
In step S309, it is determined whether the search result is correct. In the case of <last name + post name>, the search result is rarely wrong, but in the case of only <post name>, the search result often has errors. For example, in the example shown in FIG.
If “(Takeshita) Prime Minister” is a search term, (1) is OK, but (2) is incorrect. Therefore, a determination rule relating to the title is entered in the title determination rule holding unit 111, and it is determined whether the search result is correct. This rule is, for example, "It is not correct that a person's name other than the search term <surname> comes before the post name." The result determined to be incorrect is deleted from the search result holding unit 112. Then, control goes to a step S310.

In step S310, it is checked whether or not a document remains, and if so, the process returns to step S303 to repeat the processing. If not, the process ends.

Next, the present embodiment will be further described with reference to an actual example. Suppose that "Noboru Takeshita" is given as a search word. Post history database 10 corresponding to this search term
5 is as shown in FIG. Therefore, this is held in the post history holding unit 106. Also, the initial state of the search word holding unit 103 is as shown in (1) of FIG.

When the first document is taken out, the date is 1
Suppose it was January 1, 987. This is the search term holder 1
03 “1958/05/20” described in 03
, The position title corresponding to this date is checked by the position history holding unit 106 to obtain the “secretary general”. When expanding the search term, it expands to <last name + title> and <title>
If there is no synonym for "secretary general", it is expanded to three words "secretary general Takeshita", "general secretary general" and "Noboru Takeshita" as shown in FIG.
During that period, "1986/07 / 22-1987 / 11/0
5] together with the search word storage unit 103.

Thereafter, when the date of the document is until November 5, 1987, this search word can be used as it is. For example, when the date is November 6, 1987, the search word holding unit The period held at 103 will not be satisfied. Then, the position history holding unit 106 checks again to obtain the “Prime Minister”. Assuming that "Prime Minister" and "Prime" are obtained by looking up synonyms in the synonym dictionary 104, "Prime Minister Takeshita", "Prime Minister Takeshita", "Prime Minister Takeshita""PrimeMinister" is performed as shown in FIG. "Prime Minister""PrimeMinister""NoboruTakeshita"
And changes the content of the search term holding unit 103.

(Other Embodiments) In the above embodiment, the case of searching for a person name accompanied by a position name has been described.However, when searching for a name that changes itself such as a country name or an organization name, a history of each name is stored instead of the position history database. A database may be prepared, and the search term may be switched depending on the period.

2. In the above embodiment, the case where one search word is used has been described. However, even when there are a plurality of search words, a plurality of search words may be held in the search word holding unit, and the same processing may be performed for each word.

3. In the above embodiment, a case where two or more positions are not simultaneously described has been described.However, in a case where a plurality of positions are simultaneously taken, a search term is changed every time a date at which any position changes is reached. I just need.

4. In the above embodiment, the case where the documents are stored in the document database in chronological order has been described.
If the documents are not stored in chronological order, the same can be implemented by switching the search term according to the date of each document.

5. In the above embodiment, the search term is switched when the date of the name change and the date of the document match. However, when there is a delay in the document (for example, weekly magazine or monthly magazine), the delay is taken into account. May be used to switch the search term.

[0032]

As described above, according to the present invention,
When performing a search for proper nouns, it is expected that a target document can be obtained without worrying about a change in the name due to the time. Therefore, the search accuracy of the document search device is improved.

[Brief description of the drawings]

FIG. 1 is a diagram showing a basic configuration of an apparatus according to an embodiment of the present invention.

FIG. 2 is a block diagram illustrating a system configuration according to an embodiment of the present invention.

FIG. 3 is a flowchart illustrating a search processing procedure according to an embodiment of the present invention.

FIG. 4 is a diagram illustrating an example of a post history database and post history holding unit contents.

FIG. 5 is a diagram illustrating an example of an initial state of a search word holding unit;

FIG. 6 is a diagram illustrating an example of contents of a search term holding unit.

FIG. 7 is a diagram illustrating an example of contents of a search term holding unit.

FIG. 8 is an actual example for explaining a search result.

[Explanation of symbols]

 101 search word input unit 102 search word expansion unit 103 search word holding unit 104 synonym dictionary 105 post history database 106 post history storage unit 107 search processing unit 108 document database 109 document date judgment unit 110 search result judgment unit 111 post judgment rule holding Part 112 Search result holding part

────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Minoru Fujita 3-30-2 Shimomaruko, Ota-ku, Tokyo Inside Canon Inc. (56) References JP-A-4-350768 (JP, A) JP-A-5 −28199 (JP, A) Takagi, Yasuda, Shimazaki, Ikehara, “Examination of proper noun existence test method in Japanese processing”, Proc. Of the 35th IPSJ Annual Conference (late 1987) , No. 2, September 28, 1988, p. 1293-1294 (58) Field surveyed (Int. Cl. ⁷ , DB name) G06F 17/30 G06F 12/00 520 JICST file (JOIS)

Claims (3)

(57) [Claims] 1. A document storage means for storing a document together with date information for the document; a name history storage means for storing a history of names of objects in association with each name in use of the name; Based on input means for inputting, a search term input from the input means, and date information of a search target document stored in the document storage means, referencing the name history storage means and expanding the search term An information search apparatus comprising: a search word expanding unit that performs search; and a search unit that searches for the search target document based on the search word expanded by the search word expanding unit. 2. The document search apparatus according to claim 1, wherein the name history storage means stores the title of the person and the period of use as the history of the name of the person. 3. A document retrieval method for retrieving a document in a document file which stores the document together with date information on the document using a retrieval term, wherein a history of a name of an object is stored in each file by a use period of the name. A search term is input in association with the name history memory, and the name search memory is referred to based on the input search term and the date information of the search target document stored in the document file. And expanding the search term, and searching the search target document by the expanded search term.