JP3477822B2 - Document registration search system - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a registration / retrieval system for registering or retrieving a document by using a keyword, and in particular, it eliminates the influence of fluctuations in the notation of the keyword and locates the keyword in the retrieved document. Technology for detecting

[0002]

2. Description of the Related Art Conventionally, a document is registered by extracting a keyword from the document and associating the keyword with the name or identifier (ID) of the document and registering it as an index. There has been a document registration and retrieval system that searches for a corresponding document by searching the index using a keyword. Here, as a characteristic of Japanese, the katakana notation is not uniform, for example, "user" and "user", "gesture" and "gesture", "interface" and "interface", etc. There are several similar notations for expressing the same word. Therefore, when a keyword specified by the user is used as it is during the search, a document including a keyword whose expression is different from this keyword is omitted from the search.

Therefore, in Japanese Patent Laid-Open No. 62-11932, in a keyword search, an inputted keyword is transformed into a keyword having a different expression such as a character type, a notation, a kana, etc., and these keywords are used. The technique of searching by is proposed. In addition, JP-A-63-21
Japanese Patent No. 1023 proposes a technique of converting a katakana expression keyword into a vowel of a long sound, converting the expression according to a certain rule, and registering and retrieving a document using the unified expression of the keyword. There is. According to these techniques, the omission of the search due to the fluctuation of the notation of the keyword can be eliminated to some extent.

[0004]

On the other hand, in the document retrieval work, the omission of the above-mentioned retrieval is eliminated, and at the same time, it is detected where the keyword designated by the user is located in the retrieved document. This is an important issue for improving the accuracy and workability of search. In order to detect the position of such a keyword, "Salton,
G.1989. Automatic Text Processing. Reading, Mass .:
Addison-Wesley, pp.236-238 ", index the sentence position in the document of the keyword,
There has been proposed a technique capable of detecting in which sentence a keyword is located in a searched document. However, in this technology, since information about the position of the keyword is registered in the index,
The size of the index becomes enormous, which causes a problem that the process for updating the index cannot be performed quickly.

The present invention has been made in view of the above-described conventional circumstances, and an object thereof is to eliminate omission of search due to fluctuations in the notation of keywords in keyword search and to detect the position of the keyword in the searched document. To do. That is, the invention according to claim 1 converts the notation of the keyword for document registration and the notation of the keyword for document search into a unified standard notation, eliminates the omission of search due to fluctuations in the notation of the keyword, and retrieves the retrieved document. It is an object of the present invention to provide a document registration / retrieval system capable of detecting the position of a keyword in a document. The invention described in claim 2 is
An object of the present invention is to provide a device for creating search information that can eliminate a search omission due to fluctuations in keyword notation and detect the position of a keyword in a searched document in the keyword search. Another object of the present invention is to provide a device for eliminating a search omission due to fluctuations in the notation of a keyword and performing a keyword search for detecting the position of a keyword in a searched document. And The invention described in claim 4 is
By converting the notation of the keyword for document registration and the notation of the keyword for document retrieval into a unified standard notation, it is possible to eliminate the omission of retrieval due to fluctuations in the notation of the keyword and detect the position of the keyword in the retrieved document. The purpose is to provide a method for registering and searching documents that can be registered.

[0006]

A document registration / retrieval system according to a first aspect of the present invention includes a keyword extracting unit for extracting a registration keyword from a document to be registered, a standard notation and a plurality of units.
According to a predetermined rule for associating with the original notation, the registered keyword notation converting means for converting the extracted registered notation keyword of the original notation into the standard notation , and the notation associated with the original notation of the registered keyword and the converted standard notation Notation storage means for recording as correspondence information, the converted standard notation of the registered keyword, and the registration
A document retrieval storage means for storing as search information in association with information of a document to be recorded, a search keyword expression conversion means for converting the search keyword specified for document search in standard notation in accordance with the predetermined rule, Search key
Searching means for searching the corresponding document by referring to the search information by using the converted standard notation of the word; and the search key.
A notation extracting means for obtaining a corresponding original notation by referring to the notation correspondence information using a standard notation in which a word has been converted, and an original notation obtained by the notation extracting means
To the document retrieved by the retrieval means to identify the position where the keyword appears in the document.

The search information creating device according to claim 2 is
Correlate the standard notation and multiple original notations with keyword extraction means that extracts keywords from the document to be registered
According to a predetermined rule, a notation conversion unit that converts the extracted keyword of the original notation into a standard notation and a notation storage unit that records the original notation of the keyword and the converted standard notation in association with each other as notation correspondence information. And the converted standard notation of the keyword and the registered
And a document search storage unit for recording the search information in association with the information of the document to be searched .
Word to standard notation according to the rules given above
For the document search using the standard notation of the search keyword
Corresponding document by referring to the search information recorded in the storage means
And use the standard notation of the search keyword
Respond by referring to the notation correspondence information recorded in the storage means
Ask for the original notation, and put the requested original notation in front
If the keyword is matched with the searched document,
It is characterized by creating information for performing the process of identifying the position appearing in the inside .

The document retrieval apparatus according to claim 3 associates the original notation of the registered keyword extracted from the registered document with the standard notation and the plurality of original notations.
A notation storage unit that records the notation correspondence information by associating the standard notation obtained by converting the registration keyword according to a predetermined rule, the standard notation obtained by converting the extracted registration keyword according to the predetermined rule, and the registered document. Document search storage means in which information is associated and recorded as search information, notation conversion means for converting a search keyword specified for document search into standard notation according to the predetermined rule, and the standard notated search Search means for searching the corresponding document by referring to the search information by using a keyword, and notation extracting means for obtaining the corresponding original notation by referring to the notation corresponding information by using the search keyword in standard notation, The search keyword in the original notation obtained by the notation extraction means is collated with the retrieved document to find the keyword. Characterized in that it has a checking means for specifying a position which appears in the document in the.

The document registration / retrieval method according to claim 4 is a document registration / retrieval method in which a document is registered together with information necessary for retrieving the document, and the registered document is retrieved using a keyword. Document registration search system
But target the keywords extracted from the document to be registered
Prescribed rules that associate semi-notation with multiple original notations
Converting to standard notation according to the rules , recording the registered notation original information and standard notation as the notation correspondence information, and associating the standard notation of the registered keyword with the registered document information Recording the search information, converting the search keyword specified for document search into standard notation according to the predetermined rule, and referring to the search information using the standardized search keyword A document to be searched for, a standard notation of the search keyword is used to refer to the notation corresponding information to obtain a corresponding original notation, and a search keyword of the original notation is collated with the searched document Identifying the position where the keyword appears in the document,
It is characterized by executing .

[0010]

According to the document registration / retrieval system according to the first aspect, the registration / retrieval method according to the fourth aspect can be implemented as follows. That is, when registering a document, the registration keyword of the original notation extracted from the document to be registered is converted into the standard notation by the registration keyword notation conversion means, and the original notation of the registration keyword and the standard notation are associated with each other. The information is recorded in the notation storage means, and the standard notation of the registered keyword and the information of the document to be registered are associated with each other and recorded as search information in the document search storage means. Then, in the document search, the search keyword in the original notation designated for the document search is converted into the standard notation by the search keyword notation converting means, and the search information is obtained by using the standard notation of the search keyword in the searching means. The corresponding document is retrieved by referring to it, and the notation extracting means uses the standard notation of the search keyword to refer to the notation correspondence information to obtain the corresponding original notation, and the collating means retrieves the keyword of this original notation. The position where the keyword appears in the document is specified by collating with the document.

According to the search information creating apparatus of the second aspect, the registered keyword in the original notation extracted from the document to be registered is converted into the standard notation by the notation converting means,
The original notation of the registered keyword and the standard notation are associated with each other and recorded in the notation storage means as notation correspondence information. Also, the standard notation of the registered keyword and the information of the document to be registered are associated with each other as search information for document retrieval. Record in storage means. As a result, the standard notation can be used to eliminate the fluctuation of the keywords, and the information for document search that enables the identification of the appearance position of the keywords in the document by collating the original notation is created.

According to the document search device of the third aspect, the search keyword in the original notation designated for the document search is converted into the standard notation by the notation conversion means, and the standard notation of the search keyword is converted by the search means. The search information is used to search for the corresponding document, and the notation extraction unit uses the standard notation of the search keyword to refer to the notation correspondence information to find the corresponding original notation, and the collation unit determines the original notation. The search keyword is compared with the retrieved document to identify the position where the keyword appears in the document. Thereby, the original notation of the keyword extracted from the document and the standard notation are recorded in association with each other, and the document is retrieved from the document registration system in which the standard notation of the keyword and the information of the registered document are recorded in association with each other. .

Here, the present invention is particularly effective when used for katakana-represented keywords, but there are various kanji-kana-mixed expressions in which there are various Kana-kana expressions, Roman keywords in which there are various expressions, etc. It can be widely applied to the keywords of expressions in which the fluctuation of the notation exists. Further, various rules can be set for converting the original notation into the standard notation, and the standard notation does not mean a specific notation but may be a notation that is always unified according to this rule. Further, the above-mentioned document information is not only information for identifying each document such as a document name, a document identifier, an address of a recording device that stores the document, but if the size of the document search storage means permits, It may be information of the document itself.

[0014]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention, in which the influence of shaking caused by Katakana keywords is eliminated, will be described below with reference to the drawings. The registration search system is composed of a search information creation device and a document search device, and the registration search method is implemented by operating this system. FIG. 1 shows the configuration of the search information creation device, and FIG. 2 shows the configuration of the document search device.

First, the retrieval information generating apparatus is represented by a keyword extracting means 1 for extracting a keyword by performing a morphological analysis on a text of a document to be registered, and a keyword which should be converted into a standard notation, that is, katakana. Katakana comparison means 2 for judging whether the word is a word, standard notation conversion means 3 for converting a katakana keyword into a standard notation according to a predetermined rule, and notation when the keyword is present in the text (original notation) And a standard notation converted according to the rule are recorded in association with each other.
And an index registration unit 5 that performs registration processing by associating the standardized Katakana keywords or keywords other than Katakana with the document name to be registered, and an index that records the keyword and the document name to be registered in association with each other. It has a storage means 6.

The standard notation converting means 3 has a conversion rule as shown in FIGS. 5 and 6 as a table, and converts the Katakana keyword into the standard notation using this rule. For example, a keyword whose original notation is “gesture” is abbreviated to “gesture” according to the rule shown in FIG. 5, and further is subjected to substitution conversion according to the rule shown in FIG. 6 to be converted into the standard notation “gesture”. The notation storage unit 4 stores the original notation of the Katakana keyword and the standard notation in association with each other in the standard notation corresponding original notation table as shown in FIG. 4, and, for example, the original that is converted into the standard notation “gesture”. When there are a plurality of notations such as “gesture” and “gesture”, these original notations are collectively associated with the standard notation. The index storage means 6 records an index as shown in FIG. 3 as information for document retrieval, and in this index, the standard written Katakana keywords or keywords other than Katakana words such as Kanji words are registered. The document names are collectively associated and registered. For example, it is recorded that the Katakana keywords converted into the standard notation "gesture" are present in documents named "text 3" and "text 5".

Next, the document retrieval apparatus shown in FIG. 2 is, for example, a keyword extraction means 1 for performing morphological analysis on a retrieval formula input in the form of a sentence to extract a keyword, and the extracted keyword is expressed in katakana. Katakana word comparing means 2 for judging whether the word is a word or not, standard notation converting means 3 for converting the katakana keyword into standard notation according to the rules shown in FIGS. 5 and 6, and the original notation of the keyword and the rule A notation storage unit 4 for recording the standard notation in association with each other, an index storage unit 6 for recording the standard written Katakana keyword or a keyword other than Katakana and a document name to be registered, and the keyword standard. Original notation extraction means 7 for finding the corresponding original notation from the notation storage means 4 using the notation, and standardized keys Index search unit 8 to search for the corresponding document from the index storage unit 6 using over de
And a character string collating means 9 for collating the keyword in the original notation with the retrieved document to identify the position where the keyword appears in the document, and a display means 10 for displaying the keyword and the position in the document. have. In addition,
In the present embodiment, the document retrieval device and the retrieval information creation device include a keyword extraction unit 1, a Katakana word comparison unit 2, a standard notation conversion unit 3, a notation storage unit 4, and an index storage unit 6.
Is shared, and each means is reasonably utilized as the entire registration search system.

Next, in the registration / retrieval system having the above-described configuration, the process of registering a document in the index and the process of retrieving a document from the index performed by the search information creating device will be described. In the following,
A document having the document name "text 1" having the content shown in FIG. 7 will be described as a processing target. The registration process is performed according to the procedure shown in FIG. 8. First, the text portion of the document to be registered is recorded in a storage device or the like and registered (step S1), this text portion is extracted, and the keyword extraction means 1 performs morphological analysis. And a keyword is extracted from this text portion (step S2). From the text shown in FIG. 7, for example, "katakana", "notation", "interface",
Words such as “interface”, “gesture”, and “gesture” are extracted as keywords. Then, the katakana word comparing means 2 determines whether or not the extracted keywords have katakana words (step S3). If there are only keywords other than katakana words such as kanji words, the process moves to step S6 and later, which will be described later, and katakana words. If there is, the standard notation converting means 3 converts the keyword of this katakana expression into the standard notation (step S4), and further, the registration processing of this keyword in the notation storing means 4 is performed (step S).
5). In the above example, the original notation is "Katakana",
Katakana keywords such as “interface”, “interface”, “gesture”, and “gesture” are converted into standard expressions of “katakana”, “interface”, “interface”, “gesture”, and “gesture”, respectively.

The notation registration processing in step S5 is performed by the notation storage means 4 in the procedure shown in FIG. 9. First, is the standard notation of the keyword already registered in the standard notation correspondence original notation table shown in FIG. (Step S21), and if not registered, the standard notation corresponding to the keyword and the corresponding original notation are newly registered in the standard notation corresponding original notation table (Step S22). On the other hand, if the original notation corresponding to this standard notation has already been registered in the standard notation corresponding original notation table if registered (step S23), and the corresponding original notation has not yet been registered. , This original notation is additionally registered in the standard notation corresponding original notation table (step S2).
4). For example, when the standard notation “interface” is not registered in the standard notation corresponding original notation table shown in FIG. 4, the original notation “interface” corresponding to this standard notation “interface” is newly registered. If only the original notation "interface" is registered in the standard notation corresponding original notation table corresponding to the standard notation "interface", the keyword "interface" is converted to the standard notation "interface" in the original notation. And this original notation "interface" is the standard notation "interface"
Will be additionally registered corresponding to.

Returning again to the procedure shown in FIG. 8, the index registration means 5 examines the recorded contents of the index storage means 6 to find the standard written Katakana keywords or keywords other than Katakana words such as Kanji words shown in FIG. It is determined whether or not the keywords are already registered in the index (step S6). If these keywords are not yet registered in the index, the keyword and the corresponding document name are newly registered in the index (step S7). . On the other hand, when these keywords are already registered in the index, it is determined whether the document name corresponding to this keyword is registered in the index (step S
8) If it is registered, the process directly proceeds to the next process, but if it is not registered, the document name is additionally registered in the index corresponding to this keyword (step S).
9). For example, if the standard notation keyword “interface” is not registered in the index shown in FIG. 3, the document name “text 1” corresponding to the standard notation “interface” of this keyword is newly registered in the index. Further, for example, if only “text 3” is registered in the index in correspondence with the standard notation keyword “interface”, the document name “text 1” corresponding to the standard notation “interface” in this keyword is indexed. Add to.

After performing the series of processes described above, it is judged whether the keyword extracting means 1 has extracted all the keywords from the text portion of the document to be registered (step S1).
0) If the extraction of keywords has not been completed, the processing from step S2 is repeated, and the registration processing is ended when the processing is completed for all the keywords. As a result, as shown in FIG. 4, the original notation table corresponding to the standard notation of the notation storage means 4 stores the original notation of the keyword extracted from each registered document and the standard notation in correspondence, and also shown in FIG. As described above, the standard notation of the keyword corresponding to each registered document name is recorded in the index of the index storage means 6, and this information can be used for the document search described later.

Next, the document retrieval process is performed according to the procedure shown in FIG. 10. First, a retrieval formula such as a sentence designated by the user is input (step S31), and the keyword extraction means 1 uses the morpheme from this retrieval formula. Analysis is performed to extract keywords (step S32). In the following description, it is assumed that the original expression keyword “interface” is input as the search expression. Then, the katakana word comparing means 2 determines whether or not the extracted keyword has katakana word (step S33). If the keyword is not katakana word, the process proceeds to step S35 and subsequent steps, which will be described later. The standard notation converting means 3 converts the katakana expression keyword into the standard notation (step S34). In this example, the keyword in the original notation "interface" is converted into the standard notation "interface".

Next, the index search means 8 searches the index recorded in the index storage means 6 to find a document corresponding to the standard written Katakana keyword or a keyword other than Katakana (step S35). For all retrieved documents,
The appearance position of the keyword in the document is searched (steps S36 and S37), and the search process ends. In this example, 2 of "text 1" and "text 3" are associated with the standard notation "interface" from the index shown in FIG.
One document is searched.

The keyword position search processing in step S37 is performed by the original notation extracting means 7 and the character string collating means 9 in the procedure shown in FIG. 11. First, the character string collating means 9 refers to the table of the notation storing means 4. Then, it is determined whether the keyword used for the search is the standard notation (step S
41) If the standard notation is not used (that is, a search keyword other than katakana such as Kanji), the count number (remaining search number) of the keyword to be position-searched is set to "1" and the processing in step S45 and later described later I do. On the other hand, when the search keyword is the standard notation, the original notation extracting means 7 extracts the original notation corresponding to the keyword of this standard notation from the table of the notation storing means 4 (step S43), and the extracted original notation Is set as the number of remaining searches (step S44). In this example, two original notations “interface” and “interface” corresponding to the standard notation “interface” are extracted from the standard notation corresponding original notation table shown in FIG. 4, and the number of remaining searches is set to “2”. To be done. In addition,
In the case of a search keyword that is not a standard notation, there is no expansion to the original notation, and since the target for position search is this one search keyword, the number of remaining searches is uniformly set to "1" as described above.

Next, the character string collating means 9 collates the extracted original notation keywords for the Katakana keywords and the unchanged keywords for the non-Katakana keywords with the character string against the retrieved document. (Step S45) These keywords and their appearance positions in the document are highlighted and displayed on the display means 10 (step S46). In this example,
The search keyword in the original notation "interface" is expanded into two original notations "interface" and "interface" through the standard notation "interface", and as shown in FIG. 12, the appearance positions of the keywords in these original notations. Is displayed in the retrieved document.

By reducing the number of remaining searches by one, the above-mentioned processing from step S45 is sequentially performed for the keyword to be searched for position (step S47).
When the number of remaining searches is "0" and there are no keywords to be searched for a position, the position search process is terminated (step S48). As described above, according to the present embodiment, it is possible to prevent the omission of the search due to the fluctuation of the notation of the Katakana keyword by using the standard notation, and by performing the character string matching with the corresponding keyword of the original notation. The position of the keyword in the searched document can be detected, and the search reliability and operability are improved.

Here, considering a normal document, it is considered that the katakana words used in the same document are consistent in writing. For example, it is unlikely that the notation that indicates the same word "interface" and "interface" are used in the same document. For this reason, converting all the Katakana keywords into standard notation is not efficient because the same notation keywords are repeatedly converted as many times as that number. Therefore, another object of the second embodiment of the present invention described below is to eliminate the waste of repeatedly converting the katakana words of the same notation into the standard notation and achieve the improvement of the processing efficiency as a whole.

FIG. 13 shows the configuration of the search information creating apparatus according to the second embodiment of the present invention. The description of the present embodiment will be made by citing the first embodiment described above, and the same portions will be denoted by the same reference symbols and redundant description will be omitted. The search information creation device of the present embodiment is the same as that of the first embodiment described above, the keyword extraction means 1, the standard notation conversion means 3,
It has a notation storage means 4, an index registration means 5, and an index storage means 6, and as means for fulfilling the functions of the katakana word comparison means 2 of the above-mentioned embodiment, katakana discrimination means 21 and katakana notation comparison means. 22 and 22.

The katakana discriminating means 21 discriminates whether or not the keyword extracted by the keyword extracting means 1 is a word to be converted into standard notation, that is, whether or not it is a katakana word. The transmitted Katakana keywords are sent to the Katakana notation comparison unit 22. The katakana notation comparison means 22 compares with the standard notation corresponding original notation table (FIG. 4) recorded in the notation storage means 4 to determine whether the sent katakana keyword has already been converted to the standard notation. To do. That is, if the sent keyword is in the original notation of the standard notation corresponding original notation table, it can be determined that the conversion to the standard notation has already been made, and if not, it can be determined that the conversion to the standard notation has not been made yet. it can. If the katakana notation comparing means 22 determines that the katakana word is a new katakana word that has not been converted to the standard notation yet, it sends this keyword to the standard notation converting means 3 and the katakana word already converted to the standard notation. When it is determined that the corresponding standard notation is taken out from the notation storage means 4, it is sent to the index registration means 5.

Next, the process of registering a document in the index, which is performed by the search information creating apparatus of the present embodiment, will be described according to the procedure shown in FIG. First, the text portion of the document to be registered is recorded in a storage device or the like (step S51), this text portion is extracted, morphological analysis is performed by the keyword extracting means 1, and a keyword is extracted from this text portion (step S52). . Next, the katakana discriminating means 21 judges whether the extracted keyword is katakana word (step S53), and when the keyword is a word other than katakana word, it is registered in the index of the index storage means 6 in association with the document name (step S54). The process moves to step S59 and later described below. On the other hand, when the keyword is Katakana, it is determined whether or not this keyword has already been converted into the standard notation by the katakana notation comparison unit 22 and registered in the notation storage unit 4 (step S55). As a result, if the standard notation has already been converted, the corresponding standard notation is retrieved from the notation storage means 4 and the name of the document is registered in the index using this standard notation (step S56). For example, if a document named "text 6" is registered and "interface" is extracted as a keyword of the original notation from this document, "the original notation" interface "is the standard notation from the table shown in FIG. Since you can see that it has already been converted to "interface",
The document name "text 6" is additionally registered in the index shown in FIG. 3 by using this standard notation "interface" without performing conversion again.

On the other hand, as a result of the above judgment, when the keyword is not already converted into the standard notation, this keyword is sent to the standard notation converting means 3 to be converted into the standard notation (step S57), and this standard notation is used. Is used to register the name of the document in the index (step S58).
That is, the keyword is processed in the standard conversion only when it is in Katakana and the conversion into the standard notation has not been performed yet, and the wasteful process is eliminated as a whole to improve the efficiency. After the series of processes described above, it is confirmed whether the extraction of all the keywords to be extracted by the keyword extracting means 1 is completed (step S59), and further it is confirmed whether the processes for all the documents to be registered are completed. (Step S60), the registration process ends.

FIG. 15 shows the processing procedure of the keyword position search according to the third embodiment of the present invention. This embodiment does not detect the appearance position in the document for all the original notations developed from the standard notation, but detects the appearance position only for the original notation specified by the user, thereby performing unnecessary search processing. The purpose is to improve the efficiency of processing by eliminating them. The description of the present embodiment will be made by citing the first embodiment described above, and the same portions will be denoted by the same reference symbols and redundant description will be omitted.

This keyword position retrieval processing is, in addition to the original notation extracting means 7 and the character string collating means 9 shown in FIG. 2, an input means (not shown) for inputting an instruction from the user to the character string collating means 9. Done by First, the character string collating means 9 refers to the table of the notation storing means 4 to determine whether or not the keyword used for the search is the standard notation (step S71).
Processing after 74 is performed. On the other hand, if the search keyword is the standard notation, the original notation extracting means 7 extracts the original notation corresponding to the keyword of the standard notation from the table of the notation storing means 4 (step S72).
All the extracted original notations are displayed on the display means 10, and an instruction from the user is received as to which original notation the position search should be performed on (step S73). For example, when two original notations “interface” and “interface” corresponding to the standard notation “interface” are extracted from the standard notation corresponding original notation table shown in FIG. 4, which one of these original notations is located Search or get instructions from the user.

Next, the character string collating means 9 searches the retrieved document for the original notation keyword specified by the user for the Katakana word keywords and the original keyword for the keywords other than Katakana words. Collation is performed (step S74), the keyword and its appearance position in the document are highlighted and displayed on the display means 10 (step S75), and the position search process is terminated. As described above, according to the present embodiment, the appearance position is detected only for the original notation specified by the user among the original notations expanded from the standard notation,
This eliminates unnecessary search processing and improves processing efficiency.

Although the original notation for all registered documents is stored in the notation storage means 4 shown in the above embodiment, the original notation is recorded in correspondence with the standard notation for each document. You may do it. Moreover, in the above-mentioned embodiment, the registered keyword and the search keyword are converted into the standard notation by one notation conversion means 3, but the registered keyword and the search keyword are converted into the standard notation by different notation conversion means. It may be configured as follows. Further, in the above-described embodiment, the document information is recorded as the index by recording the document name in the document search recording means 6, but it is also possible to record the text information of the document.

[0036]

As described in detail above, according to the document registration / retrieval system according to the first aspect, the keyword of the original notation is converted into the standard notation by the notation converting means,
The standard notation of the keyword and the original notation are associated with each other and recorded in the notation storage means, and the standard notation of the keyword and the information of the registered document are associated with each other and recorded in the document retrieval storage means. Since the corresponding document is searched from the document search storage means by using the standard notation of the keyword, the document searched in the original notation from the notation storage means is collated with the character string to search its appearance position. It is possible to eliminate the omission of the search due to the fluctuation of the notation of the keyword and to detect the position of the keyword in the searched document, so that the reliability and operation of the search can be performed without the need for a document search storage means of a huge size. It is possible to improve the sex. Further, according to the search information creating device of the second aspect, the keyword in the original notation is converted into the standard notation by the notation converting means, and the standard notation of the keyword and the original notation are associated and recorded in the notation storing means. Since the standard notation of the keyword and the information of the registered document are recorded in the document search storage means in association with each other, the keyword notation is fluctuated without requiring the document search storage means of a huge size. It is possible to create search information that can eliminate omission of search and can detect the position of a keyword in a searched document. Further, according to the document search device of the third aspect, the keyword of the original notation is converted into the standard notation by the notation converting means, and the information of the document and the standard notation of the keyword are associated with each other by using the standard notated keyword. Search the document from the records, find the corresponding original notation from the record that associates the original notation of the keyword with the standard notation using the standard notation of the keyword, and match the keyword of the original notation with the retrieved document Since the keyword is searched for the appearance position in the document, it is possible to eliminate the omission of the search due to the fluctuation of the notation of the keyword and to detect the position of the keyword in the searched document,
The reliability and operability of the search can be improved without requiring a huge-size document search storage means.
According to the document registration / retrieval method described in claim 4, the keyword in the original notation is converted into the standard notation,
The standard notation of the keyword and the original notation are recorded in association with each other, the standard notation of the keyword and the information of the registered document are recorded in association with each other, and the corresponding document is searched using the standard notation of the keyword, Since the documents searched in the original notation using the standard notation of the keyword are collated and the appearance position is searched,
It is possible to eliminate the omission of the search due to the fluctuation of the notation of the keyword and to detect the position of the keyword in the searched document, and the reliability and operation of the search can be performed without the need for a huge memory storage for searching the document. It is possible to improve the sex.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a search information creation device according to a first embodiment of the present invention.

FIG. 2 is a block diagram showing the configuration of a document search device according to the first embodiment of the present invention.

FIG. 3 is a conceptual diagram illustrating contents of an index.

FIG. 4 is a conceptual diagram illustrating contents of an original notation table corresponding to standard notation.

FIG. 5 is a conceptual diagram illustrating the contents of a conversion rule table for standard notation.

FIG. 6 is a conceptual diagram illustrating contents of a conversion rule table for standard notation.

FIG. 7 is a diagram showing an example of text.

FIG. 8 is a flowchart illustrating a document registration processing procedure.

FIG. 9 is a flowchart showing a procedure of registration processing of an original notation table corresponding to standard notation.

FIG. 10 is a flowchart illustrating a document search processing procedure.

FIG. 11 is a flowchart showing a processing procedure of keyword position search.

FIG. 12 is a diagram showing an example of a document search result.

FIG. 13 is a block diagram showing a configuration of a search information creation device according to a second embodiment of the present invention.

FIG. 14 is a flowchart showing a document registration processing procedure according to the second embodiment of the present invention.

FIG. 15 is a flowchart showing a processing procedure of a keyword position search according to the third embodiment of the present invention.

[Explanation of symbols]

4 notation storage means 6 Document Retrieval Storage Means (Index Storage Means) 7 Notation extraction means 8 Search method 9 Verification means

─────────────────────────────────────────────────── ─── Continued Front Page (72) Inventor Kazuo Aihara 3-2-1 Sakado, Takatsu-ku, Kawasaki City, Kanagawa Prefecture KSP R & D Business Park Building Fuji Xerox Co., Ltd. (72) Inventor Tatsuomi Kita, Kawasaki City, Kanagawa Prefecture 3-2-1, Sakado, Takatsu-ku KSP R & D Business Park Building Fuji Xerox Co., Ltd. (72) Inventor Hiroshi Yamaguchi 3-2-1, Sakado, Takatsu-ku, Kawasaki City, Kanagawa Prefecture KSP R & D Business Park Building Fuji Xerox Co., Ltd. (56) References JP-A 63-211023 (JP, A) JP-A 5-274314 (JP, A) JP-A 61-267828 (JP, A) (58) Fields investigated (Int. Cl. ⁷ , DB name) G06F 17/30-17/30 419

Claims (4)

(57) [Claims] 1. A keyword extracting means for extracting a registered keyword from a document to be registered, and a predetermined notation and a predetermined notation for associating them with each other.
In accordance with a rule, a registered keyword notation converting means for converting the extracted registered keyword of the original notation into a standard notation, and a standard converted from the original notation of the registered keyword.
And notation storage means for recording a representation corresponding information in association with the notation, to the registration and converted standard representation of the keywords
That the document retrieval storage means for storing as search information in association with information of a document, a search keyword expression conversion means for converting the search keyword specified for document search in standard notation in accordance with the predetermined rule, the Searching means for searching the corresponding document by referring to the search information using the converted standard notation of the search keyword, and corresponding by referring to the notation correspondence information using the converted standard notation of the search keyword The notation extraction means for obtaining the original notation, and the original notation obtained by the notation extraction means
A document registration / retrieval system comprising: a collating unit that collates with a document retrieved by the retrieval unit to identify a position where the keyword appears in the document. 2. A keyword extracting means for extracting a keyword from a document to be registered, and a predetermined notation that associates a standard notation with a plurality of original notations.
According to a rule, a notation conversion unit that converts the extracted keyword of the original notation into a standard notation, a notation storage unit that associates the original notation of the keyword with the converted standard notation and records them as notation correspondence information, It includes a document retrieval storage means for recording the keyword converted text <br/> appended to standard representation to the registration information as search information in association, a predetermined search keyword for document retrieval According to the rules of
Convert to standard notation and change the standard notation of search keyword
Using the search information recorded in the document search storage means
Search for the corresponding document by referring to the standard of the search keyword
Notation correspondence information recorded in the notation storage means using the notation
Refer to the report for the corresponding original notation,
Match the original notation with the retrieved document
Process to identify the position where the keyword appears in the document
A search information creation device characterized by creating information for performing . 3. An original notation of a registered keyword extracted from a registered document, a standard notation and a plurality of originals
A notation storage means that records the notation correspondence information by associating the standard notation obtained by converting the registration keyword according to a predetermined rule that associates the notation with the notation, and the standard notation obtained by converting the extracted registration keyword according to the predetermined rule and the registration. a document retrieval storage means for recording the retrieval information in association with information of a document, a notation conversion means for converting the search keyword specified for document search in standard notation in accordance with the predetermined rule, the standard notation Searching means for searching the corresponding document by referring to the search information by using the searched search keyword, and notation for obtaining the corresponding original notation by referring to the notation corresponding information by using the standardized search keyword The extraction means and the original search keyword obtained by the notation extraction means are referred to the searched document. To document search device to which the keyword is characterized by having a, a verification means for identifying a location appearing in the document. 4. A document registration / retrieval method for registering a document together with information necessary for retrieving the document and retrieving the registered document by using a keyword.
But target the keywords extracted from the document to be registered
Prescribed rules that associate semi-notation with multiple original notations
Converting into standard notation according to the rules, the notation storing means of the document registration search system records the original notation of the registered keyword and the standard notation in association as notation correspondence information, and document search of the document registration search system A storage unit for storing the standard expression of the registered keyword and the information of the registered document in association with each other and recording the information as search information; and a search keyword expression conversion unit of the document registration search system.
Referring but converting the search keyword specified for document search in standard notation in accordance with the predetermined rule, the search unit of the document registration search system, the search information using the standardized notation Searches To search for a corresponding document, a notation extraction means of the document registration and retrieval system refers to the notation correspondence information using the standard notation of the search keyword to obtain a corresponding original notation, and a document registration search A method of registering and retrieving a document, characterized in that the collating means of the system collates the retrieval keyword in the original notation with the retrieved document to identify the position where the keyword appears in the document. .