JP2000020537A - Text retrieving device and computer-readable recording medium having recorded text retrieving program thereon - Google Patents

Abstract

PROBLEM TO BE SOLVED: To execute accurate text retrieval and keyword development to be inputted to the text retrieval. SOLUTION: The text retrieving device is provided with a data base(DB) retrieving condition input means 21 for specifying a table format DB and inputting a DB retrieving expression describing a condition for retrieving a required line from the table format DB, an acquired attribute defining means 25 for defining a column item having a keyword to be acquired as a value in the DB, a DB retrieving statement preparing means 22 for selecting a line satisfying the DB retrieving expression and preparing a DB retrieving statement for acquiring the keyword to be the value of the column item defined by the means 25 from the selected line, a development word extraction means 25 for executing the DB retrieving statement and acquiring the keyword to be the value of the column item defined by the means 25, and a keyword expression preparing means 24 for preparing a keyword expression by combining acquired keywords.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a text search apparatus capable of performing a text search with high precision and expanding a keyword as an input thereof, and a computer readable program storing a text search program for realizing the text search apparatus by a computer. It relates to a recording medium.

[0002]

2. Description of the Related Art Heretofore, in this type of text search apparatus, using vocabulary knowledge information such as a dictionary or a thesaurus, a keyword having a synonym, a different notation, a synonym, or a hierarchical relationship with an input keyword is used as a search keyword. There is a technique for improving the accuracy of the search by adding the information. Adding keywords in this way is generally "expanding keywords"
Say An example of a patent application of this type of technology is disclosed in
Japanese Patent Application Laid-Open No. 2-11932, Japanese Patent Application Laid-Open No. 62-248032, Japanese Patent Application Laid-Open No. 4-10062, and the like.

[0003] Japanese Patent Application Laid-Open No. Hei 4-10062 describes several methods for expanding an inputted keyword. One is to expand keywords into lower words. For example, expanding from the keyword “university” to its lower-level words “University of Tokyo” and “Osaka University”
A search formula such as OR “Todai” OR “Osaka University” is created. The other is a method of expanding a keyword into a broader word (concept), and further expanding it into lower words of the concept. For example, the keyword "Todai" is expanded to the upper term "university" on the thesaurus, and further expanded to the lower term "Osaka University". In addition, a method of expanding only vocabulary whose hierarchy depth is lower than the n-th stage or expanding and extracting all lower vocabulary up to the n-th stage is shown.

For example, when it is desired to search for a document relating to "PC", there is a case where it is desired to search for an article including a specific name such as a product name and a model name without including the keyword "PC". Using the keyword expansion technology, the keyword "PC" can be expanded to the lower-level word "notebook PC" or the product name / model name using a thesaurus containing proper nouns such as product names and model names, and searched. , It is possible to reduce search omissions.

[0005]

As a first problem of the conventional text search apparatus that performs keyword development using such a thesaurus, there is a problem that construction of the thesaurus requires labor. This is because, for example, in the case of a search request related to automobiles, by manufacturer (eg, “A company” or “B company”), by purpose (“family car”, “sports car”, etc.), or by shape (“sedan”) Search requests from various viewpoints such as “coupe” and “open”) and sizes (“small car” and “large car”) are assumed, and each viewpoint needs to be classified by a thesaurus (tree structure). The words indicating these classifications do not necessarily have a clear upper / lower relationship with each other, and a plurality of thesauruses must be constructed, which requires labor and complicates management. In this regard, the technology described in Japanese Patent Application Laid-Open No. 4-10062 does not explicitly use a thesaurus, but uses a vocabulary dictionary and a relation list indicating higher / lower relationships in the vocabulary dictionary, and uses a vocabulary dictionary and a relation list. It describes that a deep hierarchical relationship like a thesaurus can be expressed. However, managing such vocabulary dictionaries and relation lists is very time-consuming, and the time and effort required to create them is almost the same as that of a thesaurus. As a result, the first problem cannot be solved.

A second problem of the prior art is that it is difficult to handle numerical information in a thesaurus. Numerical conditions, such as price, displacement, and size, are also assumed as search requests for automobiles. For example, numerical conditions such as "2,000,000 yen or less" and "2000cc or more" are required, but numerical information is difficult to treat as a thesaurus. As a text search method capable of specifying numerical conditions, a “search system based on text structuring” (Kishimoto et al., IPSJ Transactions, Vol. 35, No. 5, pp. 908 to 916,
(1994), there is a method using an attribute dictionary. The numeric character strings in the query and the dictionary are converted to numerical values, and unit unification processing (“5000g” = “5k
g ”), and the comparison as a numerical value is performed.
Words on the thesaurus and information on the attribute dictionary are related by ID. However, it still has a plurality of dictionaries, and it can be said that there is a maintenance problem.

Further, as a third problem of the prior art,
There is almost no thesaurus at the proper noun (product name / company name) level. If there is no thesaurus including proper nouns, it is not possible to sufficiently expand it into instances.

[0008] The present invention has been made in view of the problems of the prior art, and an object of the present invention is to solve these problems.

[0009]

SUMMARY OF THE INVENTION The present invention solves the above-mentioned problems by a text search apparatus characterized by query expansion using a tabular database. The tabular database is, for example, a relational database (Relational database).
This is a tabular database such as Database (RDB), and is widely used for managing product information. The product information, for example, in the case of an automobile, is information such as product name, model name, manufacturer, type such as shape, purpose, and size, price, and displacement as described above.

[0010] By referring to a database suitable for the keyword, the keyword can be developed into a product name or the like, and the development can be realized in consideration of numerical conditions. Also, since a plurality of classification viewpoints can be managed in a single table, maintenance is easier than in a thesaurus with a tree structure. Since the RDB is widely used for database search, there is a high possibility that an existing one can be used. ADVANTAGE OF THE INVENTION According to this invention, management of deployment information becomes easy and can respond to various search requests including numerical information.

Therefore, the present invention is configured as follows.

First, a text search apparatus according to a first aspect of the present invention is a DB search condition input means capable of designating a tabular database consisting of rows and columns and inputting a database search formula for searching a text database. And an acquisition attribute defining means for defining a column item having a keyword to be acquired as a value in the tabular database, and selecting a row satisfying the database search formula from the tabular database specified by the DB search condition inputting means. A DB search statement creating means for creating a database search statement for acquiring a keyword which is a value of a column item defined by the acquisition attribute defining means from the row; and executing the database search statement to acquire the keyword. Expansion word extraction means for acquiring a keyword which is a value of a column item defined by attribute definition means; Using the keywords obtained by the extraction means, characterized in that it is configured to have at least a keyword expression creating means for creating a keyword expression to search the text database.

Further, in the first invention, the DB
Search condition input means for displaying a list of the tabular database, DB user selecting means for providing a user interface for selecting the tabular database,
DB search expression input means for obtaining a column item from the tabular database selected by the DB user selection means, displaying the list item, and providing a user interface for inputting a value for each column item. I have.

Next, a text search apparatus according to a second invention of the present invention includes a text search condition input means for inputting a keyword or a natural language sentence for text search by a user, and the text search condition input means. A database search formula that selects a tabular database consisting of appropriate rows and columns for a keyword or natural language sentence entered by the user and describes conditions for retrieving a desired row from the tabular database From the tabular database selected by the DB search condition creating unit, the acquisition attribute defining unit defining a column item having a keyword as a value to be acquired in the table format database, Select a row that satisfies the search expression, and from that row the values of the column items defined by the acquisition attribute definition means A DB search statement creating means for creating a database search statement to retrieve the keyword,
An expansion word extraction means for executing the database search statement to obtain the keyword of the column item defined by the acquisition attribute definition means, and a keyword expression for text search using the keyword obtained by the expansion word extraction means It is characterized by comprising at least a keyword expression creating means for creating.

In the second invention, the DB
A search condition creating means for storing an expanded word table storing information indicating a tabular database to which the keyword belongs and column items in the keyword tabular database; and a keyword input via the text search condition input means or DB automatic selecting means for selecting the tabular database by comparing a keyword obtained from a natural language sentence input by the text search condition input means with the developed word table, and database selection by the DB automatic selecting means. It is also characterized in that it comprises a DB search formula creating means for creating a database search formula from the used keywords and the column items in the tabular database of the keywords.

Next, a text search device according to a third invention of the present invention is provided through a text search condition input means for inputting a keyword or a natural language sentence for text search by a user, and the text search condition input means. A database search formula that selects a tabular database consisting of appropriate rows and columns for a keyword or natural language sentence entered by the user and describes conditions for retrieving a desired row from the tabular database DB search condition creating means for creating a database, acquisition attribute defining means for defining a column item having a keyword to be acquired in the tabular database as a value, and database search from the tabular database selected by the DB search condition creating means Select a row that satisfies the expression, and select a key that is the value of the column item defined by the DB search statement creation means for creating a database search statement for acquiring a word, and an expansion word for executing the database search statement and acquiring a keyword which is a value of a column item defined by the acquisition attribute definition means The extraction means and the keyword expression combining the keywords input by the text search condition input means or the keyword expression combining the keywords obtained from the natural language sentences input by the text search condition input means, and the expanded word extraction means It is characterized by comprising at least keyword expression integrating means for integrating a keyword expression obtained by combining the obtained keywords and creating a keyword expression for text search.

Further, the text search device according to a fourth invention of the present invention, in the first invention, has a keyword used for text search from a database search formula input through the DB search condition input means as a value. A keyword attribute defining means for defining a column item or a joining condition between column items in a tabular database; and a keyword for extracting a keyword from the database search formula based on the definition of the keyword attribute defining means and creating a combined keyword expression Combining means, wherein the keyword formula creating means integrates the keyword formula obtained by the keyword combining means and the keyword formula obtained by combining the keywords obtained by the expanded word extracting means, and With keyword expression integration means to create a keyword expression to search And wherein the Rukoto.

Further, the text search apparatus according to a fifth aspect of the present invention, in the first to third aspects, comprises an attribute expansion relation defining means for defining an expansion relation between column items in the tabular database. The DB search command sentence creating unit includes a database search expression input by the DB search condition input unit or a database search expression created by the DB search condition creation unit as a definition of the attribute expansion relationship definition unit. And developing a database search command sentence for acquiring a keyword which is a value of a column item defined by the acquisition attribute defining means.

The text search device according to a sixth aspect of the present invention is the text search device according to the first to third aspects, wherein the ambiguity defining a judgment criterion as to whether or not the keyword acquired by the expanded word extraction means is an ambiguous word is defined. Judgment rule definition means, removing an ambiguous word from the keyword acquired by the expanded word extraction means based on the judgment criteria defined by the ambiguous judgment rule definition means,
An ambiguous word removing means for outputting the keyword from which the ambiguous word has been removed to the keyword expression creating means or the keyword expression integrating means.

The text search device according to a seventh aspect of the present invention is the text search device according to the first to third aspects, wherein the ambiguity defining a criterion for determining whether or not the keyword acquired by the expanded word extraction means is an ambiguous word is defined. Determination rule defining means, an ambiguous word determining means for determining an ambiguous word based on a criterion defined by the ambiguous determination rule defining means from the keyword acquired by the expanded word extracting means, and the expanded word extracting means A related word defining means for defining a condition for extracting a value of a column item serving as a related word for the ambiguous word from the acquired row, and acquiring a related word for the ambiguous word according to the definition of the related word defining means. Related word acquiring means for creating a keyword expression by combining a word and a related word thereof, wherein the keyword expression creating means or the keyword expression integrating means comprises: Keywords prepared in the acquiring means be used, characterized by creating a keyword expression to search the text database.

The text search device according to an eighth aspect of the present invention, in the first to third aspects, displays all the keyword expressions created by the keyword expression creating means or the keyword expression integrating means, It is characterized by having a development result editing means for providing an interface for a user to perform editing such as addition and deletion.

A text search apparatus according to a ninth aspect of the present invention includes a text search condition input means for inputting a keyword for text search by a user, and a tabular database comprising rows and columns. A similarity determination defining means for defining a similarity condition for determining similarity between rows based on a value of a column item in a database, and a row including a keyword input by the text search condition inputting means is searched from the tabular database. A similar keyword acquiring unit that acquires, as a keyword, a value of a column item having a keyword input by a user as a value from the line and a line satisfying the similar condition, and a keyword input through the text search condition input unit. Combining the combined keyword expression with the keyword acquired by the similar keyword acquiring means It is characterized by being configured with a keyword expression integration means for creating a keyword expression and integration keyword expression at least a.

[0023]

Embodiments of the present invention will now be described in detail with reference to the drawings.

[First Invention] (Explanation of Structure) FIG. 1 is a block diagram showing the structure of a first invention of a text search apparatus according to an embodiment of the present invention.

Referring to FIG. 1, the first invention comprises a DB storage unit 20, a DB search condition input unit 21, a DB search command sentence creation unit 22, an expansion word extraction unit 23, a keyword expression creation unit 24, an acquisition attribute definition It comprises means 25, text storage means 10, text search means 11, text search formula input means 12, and search result display means 13.

The text storage unit 10, the text search unit 11, the text search expression input unit 12, and the search result display unit 13 are units that constitute an existing text search device, and are not characteristic features of the present invention. Omitted.

The DB storage means 20 stores one or more tables in the form of a table (hereinafter, DB) composed of rows and columns. As a tabular DB, for example, a relational database (Relative Database: hereinafter, R
DB) are well known. Each DB stored in the DB storage means 20 has a DB name (hereinafter, DB name).
Is given. Each DB consists of one column header row and zero or more rows with data values (reference: "Standard S
QL ", by CJ Date, p. 42, Toppan. The column heading rows consist of named columns, and these names are the attributes of each row. Each row is the attribute of each column specified in the column heading row. It has a unique value (hereinafter, attribute value) for the attribute.The DB storage means 20 is also a means for storing an existing DB, and is not a characteristic part of the present invention.

The DB search condition input means 21 is a means for inputting a DB name in the DB storage means 20 and an expression (hereinafter, a DB search expression) representing a condition of a value satisfied by an attribute in the DB. .

The DB search command sentence creating means 22 receives the DB of the DB name input by the DB search condition input means 21
This is a means for selecting a row that satisfies the DB search expression and creating a DB search command sentence for acquiring a value of a specific attribute as a keyword. This DB search statement is, for example, the S of RDB.
It is a QL sentence. The keyword acquired here is a word obtained by expanding the keyword given to the text search device by the user, and is hereinafter referred to as a developed word. Which attribute value is acquired as a development word is defined by the acquired attribute definition means 25.

The expanded word extracting means 23 is a means for executing the DB search instruction sentence created by the DB search instruction sentence creating means 22 to acquire an expanded word.

The keyword formula creating means 24 creates a keyword formula combining the expanded words obtained by the expanded word extracting means 23. For example, it is conceivable to create a keyword expression in which keywords are combined by a search operator such as AND / OR, or to create a keyword expression in which keywords weighted by a vector space model are combined.

The text search formula input means 12 inputs the keyword formula created by the keyword formula creation means 24 to the text search means 11.

The text search means 11 searches the text storage means 10 for a text which satisfies the keyword formula input by the text search formula input means 12. The search result display means 13 displays a search result of the text search means 11.

(Explanation of Operation) Regarding the operation of the text search apparatus of the first invention, the text stored in the text storage means 10 is related to an article, and the DB stored in the DB storage means 20 is "automobile". And a specific example in which the user searches for an article related to "A company or B company's sedan of 2 million yen or less" (hereinafter, this example is referred to as the first embodiment and Name.)
This will be described with reference to FIG.

In the case of this example, the DB search condition input means 21
Are explicitly input by the user through the database "automobile DB" as a DB name and "a sedan of 2 million yen or less from Company A or Company B" as a DB search formula. Hereinafter, this operation will be described.

FIG. 2 is a diagram showing an example of a specific implementation for inputting a DB name and a DB search formula by the DB search condition input means 21 in FIG. The DB search condition input unit 21 includes a DB user selection unit 210 and a DB search expression input unit 211.

The DB user selection means 210 is a means for displaying a list of DBs stored in the DB storage means 20 and providing a user interface for selecting a DB. The DB to be selected is not limited to one, but may be plural.
The DB storage means 20 is a storage device or storage medium such as a hard disk or a CD, and stores a DB prepared by a user or a DB prepared by a service provider. The DB user selection unit 210 accesses the DB storage unit 20 by specifying a device or a directory when the user has the DB storage unit 20 locally, or by specifying a server or a URL when the DB storage unit 20 is on a network. To display a DB list and select a DB according to a user's instruction. The device, directory, server, URL, and the like to be specified are defined in advance.

The DB search formula input unit 211 obtains a list of attribute names of the DB selected by the DB user selecting unit 210, presents the list to the user, and provides a user interface for inputting a value for each attribute. It is.

The DB user selecting means 210 is, for example, the one shown in FIG.
A list of DBs such as 7 is displayed, and in this example, the user selects “automobile DB” as the DB to be used for the search.

When "vehicle DB" is selected, the DB search formula input means 211 inputs values for "manufacturer", "shape", "purpose", "price" and "displacement" which are the attributes of the vehicle DB shown in FIG. Create a menu for input.
Such a menu is prepared in advance or created automatically. An example of a method of automatically creating the information will be described later. FIG. 18 is an example of an attribute value input menu created by the DB search expression input unit 211. A frame for inputting a value for each attribute is prepared with the attribute name as a heading, and each frame is provided with a button indicated by an inverted triangle (“▽”). When the button is pressed, a list of values of the attribute name is displayed as shown in FIG. 19, and the user can select one required value from the list for one frame. In FIG. 18, the attribute values for the same attribute are combined under an OR condition, and the attributes are combined under an AND condition. However, the combination condition is not limited to this.

For numerical attributes such as "price" and "displacement", the user directly inputs numerical values. To enter a numeric attribute, press the button indicated by “▽” as shown in FIG. 20 for conditions indicating a numerical range such as “greater than”, “less than”, “greater (super)”, “smaller (less than)”. You can select from the list displayed. “More” is “≧”, the following is “≦”,
If “greater (super)” is associated with “>” and “smaller (less than)” is associated with “<”, for example, “price”
By inputting a numerical value of "200000" and selecting a condition of "less than or equal to", a DB search formula of "price≤200000" can be obtained. Therefore, DB search condition input means 2
1, “automobile DB” as a DB name and “sedan of 2 million yen or less from Company A or Company B” can be input as a DB search formula.

Here, the flow of automatic menu creation in the DB search expression input means 211 will be described with reference to the flowchart of FIG. In S140, the DB search expression input unit 211 selects the DB selected by the DB user selection unit 210.
Get N attribute names. The attribute name to be obtained is a heading of the input menu in the DB search expression input unit 211, and all the attribute names are to be obtained, or which attribute name is to be obtained is determined in advance. For example, FIG.
5 "Manufacturer" and "Shape" as attribute names from the automobile DB
Acquire "purpose", "vehicle name", "price", and "displacement". S1
At 42, K (K = 1, 2,
.., N), it is determined whether the type of the attribute value is character type or numeric type. The first attribute is "Manufacturer"
If so, the determination result of S142 is a character type and the process proceeds to S143, and all the attribute values (“Company A”, “Company B”...) Are acquired. In step S144, the attribute values obtained in step S143 are sorted (sorted) in dictionary order, and unique (duplicate removal processing) is performed to obtain a non-duplicated attribute value list sorted in dictionary order. In S145, an HTML tag as shown in FIG. 22, for example, is added to the list of attribute names and attribute values. If the result of the determination in S142 is a numeric type (eg, "price"), the flow advances to S146 to create an HTML text as shown in FIG. In S147, the HTML texts created in S145 and S146 are sequentially combined. When the processing ends in S148, the HTML of FIG.
By displaying the text with a browser that supports HTML, a screen interface such as a list display button or frame as shown in FIG. 18 can be realized.

FIG. 21 shows an example of a screen on which all the attribute values are input in this embodiment. With this screen, a DB search command of “(automobile DB. Maker =“ company A ”OR“ company B ”) AND (automobile DB. Shape =“ sedan ”) AND (automobile DB. Price ≦ 200000)” is issued as a DB search instruction. It can be input to the sentence creation means 22.

The DB search statement creation means 22 is a
B, the above DB search expression "(automobile DB. Maker =
“Company A” OR “Company B”) AND (Automotive DB. Shape =
"Sedan") AND (Automobile DB. Price ≤ 200,000
0) is selected, and a DB search command sentence for acquiring an attribute value defined by the acquisition attribute defining means 25 as an expansion word is created. The acquisition attribute definition unit 25 defines in advance which attribute value is to be used as a development word, or defines the attribute value according to a user's specification when executing a search. Here, it is assumed that it is defined that the vehicle name and the model name are acquired as expansion words.

The DB search statement created by the DB search statement creation means 22 is, for example, an SQL statement in an RDB.
The format is roughly as shown in FIG. FIG.
(A) describes the DB name input by the DB search condition input unit 21 in the FROM line of the SQL sentence. WHER
In the row E, the DB search formula input by the DB search condition input unit 21 is described. In the SELECT line, an attribute name defined by the acquisition attribute definition unit 25 is specified. FIG. 16 (B)
Is a specific example of a DB search statement created by the DB search statement creation unit 22. In FIG. 16B, for convenience, DB
Are represented in a format of “DB. Attribute name”.

The expanded word extracting means 23 executes the DB search command sentence as shown in FIG. 16B created by the DB search command sentence creation means 22, and the vehicle DB shown in FIG. Name “AAAY” “BBBX”, type name “bbb”
x01 "is obtained as a development word.

The keyword expression creating means 24 combines the expansion words obtained by the expansion word extracting means 23 with, for example, an OR condition by using a keyword expression “AAAY” OR “BBBX” OR
“Bbbbx01” ”is created and input to the text search formula input means 12. Here, the combination condition of the expansion word is not limited to OR. For example, “BBBX” and “bbb”
x01 ”and the attribute values obtained from the same row are AND-connected to“ “AAAY” OR (“BBBX” AND “b
bbx01 ")" or create a keyword expression
EAR, and “AAAY” OR (“BBBX” N
EAR "bbbx01") "may be created.

The text search means 11 searches the text storage means 10 for a text satisfying the keyword formula created by the keyword formula creation means 24.

As described above, in the present embodiment, when the user executes a text search, "2 company A or B company"
Regardless of the specific name such as car name, model name, etc. for `` sedan under 100,000 yen ''
A keyword with a specific name can be given.

[Second Invention] (Explanation of Structure) FIG. 3 is a diagram showing a structure of a text search device according to a second embodiment of the present invention. Compared to the configuration of the first embodiment of the present invention, a DB search condition creating unit 42 is used instead of the DB search condition input unit 21,
A text search condition input means 26 has been added as a new means.

The text search condition input means 26 is a means for the user to input a free keyword or a natural language sentence for text search.

The DB search condition creating means 42 automatically selects the DB in the DB storage means 20 from the keywords input by the text search condition input means 26 or the words constituting the natural language sentence, and automatically executes the DB search formula. It is a means to create it.

(Explanation of Operation) The operation of the text search apparatus shown in FIG. 3 will be described with reference to a specific embodiment (hereinafter, this embodiment is referred to as a second embodiment). In the text search condition input means 26, the keyword expression “(“ Company A ”OR
An operation example in the case where “Company B”) AND “Sedan” is input will be described.

FIG. 4 is a diagram showing a specific example of the realization of the DB search condition creating means 42 in FIG. In FIG. 4, the DB search condition creation unit 42 includes a DB automatic selection unit 212, a developed word table 213, and a DB search expression creation unit 215. The expanded word table 213 indicates which D
9 is a table showing which attribute of B is an attribute. The DB automatic selection unit 212 is a unit that selects an optimum DB for the keyword input by the text search condition input unit 26 based on the contents of the developed word table 213. The DB search formula creation unit 215 is a unit that creates a DB search formula from the keywords used for the DB selection by the DB automatic selection unit 212.

Now, the keyword expression “(“ Company A ”OR“ Company B ”) AND is input via the text search condition input means 26.
When "sedan" is input, the DB automatic selection unit 212
First, the expanded word table 213 is referred to. FIG. 24 shows an example of the contents of the expanded word table 213. According to the developed word table 213, “Company A” is “automobile DB. Maker” or “home appliance DB. Maker”, “Company B” is “automobile DB. Maker” or “motorcycle DB. Maker”,
“Sedan” stores information of “automobile DB. Shape”. The DB automatic selection unit 212 determines that the DB to which these three keywords relate most is “automobile DB”.
Therefore, "automobile DB" is selected as an appropriate DB. The expanded word table 213 is stored in the DB storage unit 20.
By extracting a character-type attribute value (keyword) from each of the DBs in combination with the DB name / attribute name to which the attribute value belongs, it can be easily created mechanically.

The DB search formula creation means 215 uses the keyword used for the DB selection by the DB automatic selection means 212 and its attribute information to indicate “automobile DB.maker =“ Company A ””
"Automobile DB. Maker =" Company B "", "Automobile DB.
A shape = “sedan” ”is created as a DB search expression, and is input to the DB search command sentence creating means 22.

The subsequent operation is the same as that of the first embodiment, and the description is omitted.

Also, in the configuration of FIG.
A configuration in which the user information memory 214 is referred to as shown in FIG. The user information memory 214 is a means for storing personal information such as history information and interest information of the user using the search device.

Next, regarding the operation of the text search apparatus when the configuration of the DB search condition creation means 42 is as shown in FIG. 5, the text search condition input means 26 uses the keyword expression "" Company B "AND" Sports "". This will be described using an operation example in the case where is input.

In the developed word table 213, "Company B" stores "automobile DB. Maker" or "motorcycle DB. Maker", and "sports" stores information of "automobile DB. Purpose" or "motorcycle DB. Purpose". Have been. These two keywords are related by the same number in both “automobile DB” and “motorcycle DB”.
12 cannot be decided on one side.

Therefore, the automatic DB selection means 212 refers to the contents of the user information memory 214. FIG. 25 shows an example of the contents of the user information memory 214. FIG. 25 shows that the DB names that the user has used and the frequency of use are stored as personal information. According to the use frequency information, it is presumed that the user is interested in the most frequently used vehicle. DB automatic selection means 212
Selects the automobile DB having a high use frequency in the user information memory 214 prior to the motorcycle DB. The user information stored in the user information memory 214 explicitly specifies not only the history information such as the frequency of use of the DB but also the interest information of the user by the user. May be registered in the user information memory 214 in advance).

"Automobile DB" by the DB automatic selection means 212
The operation after selecting is the same as in the previous embodiment, and the description is omitted.

In the text search apparatus shown in FIG. 3, what is input to the text search condition input means 26 is
It can be a free keyword or a natural language. FIG. 6 is a diagram illustrating an example of a configuration of the text search condition input unit 26 for enabling input of a natural language.

In FIG. 6, the text search condition input means 26 includes a control processing means 263, a controlled word dictionary 264, a search request sentence input means 261, a sentence analysis means 262, a word dictionary 2
60, a numerical expression extracting unit 265, a numerical extraction rule defining unit 266, a numerical search formula creating unit 267, and a numerical conversion rule defining unit 268.

The search request sentence input means 261 is a means for the user to input a natural language search request sentence. The sentence analyzing unit 262 analyzes the search request sentence with reference to the word dictionary 260 and performs word division or part of speech information addition. The control processing means 263 refers to the controlled word dictionary 264 for the word string as a result of the sentence analyzing means 262 and replaces a word serving as a search keyword with a controlled word.

The numerical value extraction rule defining means 266 is a means for describing a numerical expression pattern to be extracted by focusing on a character string indicating a unit of the numerical expression. Numeric expression extraction means 265
Is a means for extracting a word string matching the pattern defined by the numerical value extraction rule definition means 266 from the word string as a result of the sentence analysis means 262 as a numerical expression together with its attribute information.

The numerical conversion rule definition means 268 is a means for defining a conversion rule from a numerical representation of a character string to a numerical value.
Numerical search expression creating means 267 includes numerical expression extracting means 265
Is a means for creating a DB search formula relating to a numerical value from the numerical expression extracted and the attribute information thereof.

The operation of the text search device shown in FIG.
A case where a natural language sentence “I want to read an article on a sedan of A or B of 2 million yen or less” will be described. Here, “A” is an abbreviation for “Company A”, and “B” is an abbreviation for “Company B”.

When this search request sentence is input by the search request sentence input means 261, the sentence analysis means 262
With reference to 60, "A / also / ha / B // 200 / 10,000 /
Yen / around / of / sedan / related / article / read / read /
Division into word strings such as "/". This type of sentence analysis method is often used (Reference: "Storage of Japanese language dictionary and automatic division of Japanese sentence", Nagao et al., Information Processing, Vol. 19, No. 6, 1978). Control processing means 26
3 is a controlled word dictionary 2 in the word string resulting from the sentence analyzing means 262.
64, "A" becomes "Company A", "B" becomes "B"
Company ". Such replacement processing is performed in order to control a keyword to a headword of the developed word table 213 when the DB automatic selection unit 212 refers to the developed word table 213. As described above, a method of controlling free keywords having the same meaning as input keywords in a text search into words that can be processed by the search device is often used, and is disclosed in Japanese Patent Application Laid-Open Nos. 5-250416 and 8-
No. 249341. FIG. 23 shows an example of the contents of the controlled word dictionary. The controlled word dictionary 264 includes a set of a keyword heading assumed as an input and its controlled word.

The DB automatic selection means 212 reads “Company A / or / ha / Company B // 200 / million / yen / around / of / sedan /
The most appropriate DB is selected by inputting the word string “// related / article / read / want /”. As for the method of selecting the DB, referring to the developed word table 213 or the user information memory 214, the keywords "A company", "B company", and "sedan" are used in the same manner as in the second embodiment. DB ”is automatically selected.

The DB search formula creation means 215 uses the keyword used for the DB selection by the DB automatic selection means 212 and its attribute information to obtain “automobile DB.maker =“ Company A ””,
"Automobile DB. Maker =" Company B "", "Automobile DB.
Shape = “sedan” ”is created as a DB search formula.

The numerical expression extracting means 265 is, for example, a “circle”.
Paying attention to the unit expression such as
From the keyword input in step 2, “200
"10,000 yen or less." In the numerical value extraction rule definition means 266, “numeric string +“ yen ”(price unit)” or “numeric string +
A pattern of a character string arrangement such as “yen” (price unit) + “more than” OR “less than or equal to”... Is described as an extraction condition, and the numerical expression extraction unit 265 defines a pattern that matches this extraction condition as a sentence. It is extracted as a numerical expression from the word string that is the result of the analysis means 262. As a result, the numerical expression extraction means 265 inputs the word string information ““ 200 ”,“ million ”,“ yen ”(price unit)“ not more than ”” and the attribute information to the numerical search expression creating means 267.

The numerical value retrieval formula creating means 267 outputs “200”, “million”, “yen” inputted from the numerical value expression extracting means 265.
The part corresponding to the numerical value in the character string information “less than or equal to” is converted into a numerical value. The numeric conversion rule definition means 268 defines, for example, numeric characters “1” “2”.
“10,000” is replaced by the numbers “1”, “2”, “3” ... “00” “0”
00 ”... Using this, the “200” and “million” portions of the character string information “200”, “million”, “yen”, and “less” are converted to a numerical value of “2,000,000”.

As shown in FIG. 6, DB search formula input means 211
By providing the above, the above-mentioned DB search formula creation means 215
In addition, the result of the numerical search formula creation means 267 is displayed on the attribute value input screen as shown in FIG.
"Company B" automatically enters "Sedan" in the "Shape" box and "2,000,000 yen or less" in the "Price" box, and provides an interface that allows the user to add and delete attribute values. It is also possible to provide. In a configuration without the DB search formula input unit 211, the DB search formula creation unit 215
And DB of the result of numerical search formula creation means 267
It will be passed to the search statement creation means.

The operation after the DB search formula is input by the DB search formula input means 211 is the same as that of the first embodiment, and the description is omitted.

[Third Invention] (Explanation of Structure) FIG. 7 is a diagram showing a structure of a third invention of a text search apparatus according to an embodiment of the present invention. As compared with FIG. 3 of the second invention, a keyword expression integrating unit 43 and a combination condition selecting unit 27 are added instead of the keyword expression creating unit 24.

The keyword formula integrating means 43 is a means for creating a keyword formula by combining a keyword formula input by the text search condition input means 26 and a keyword formula based on a combination of the expanded words obtained by the expanded word extracting means 23. .

The combination condition selection means 27 is means for the user to select a condition for combining the keyword expression input by the text search condition input means 26 and the keyword expression based on the combination of the expansion word obtained by the expansion word extraction means 23. is there. The combining condition selecting means 27 is not an essential constituent means,
If the combining condition selecting means 27 is not provided, the keyword formula integrating means 43 combines the respective keyword formulas under predetermined conditions.

The keyword formula created by the keyword formula integrating means 43 is input to the text search formula input means 12, and
Used as a keyword expression for text search in the text search means 11.

(Explanation of Operation) Regarding the operation of the text search apparatus shown in FIG. 7, the text search condition input means 26 uses the keyword expression (A) ““ A company ”AND“ sedan ””.
This will be described using a specific example (hereinafter, this example is referred to as a third example) in a case where is input.

When this keyword expression is input, "automobile DB" is selected by the search condition input means 21 in the same manner as in the second embodiment, and the expanded word extraction means 23 searches the automobile DB for the DB. Formula "(Manufacturer =" Company A ") A
ND (shape = “sedan”) ”is executed, and“ “AAX” OR “AAAY” O
The keyword expression (B) of R “aaax01” OR “aaaax02” is obtained.

The joining condition selecting means 27 is a means for the user to select the joining condition of the keyword expression (A) and the keyword expression (B). For example, whether the joining condition is joined by an OR condition or an AND condition is determined. You can make a selection. O
The R join can reduce search omission, and AND
Joining can increase the precision of a search. The user can select one according to the purpose. Assuming that the OR condition is selected, the keyword expression integrating means 43 creates a keyword expression of “keyword expression (A) OR keyword expression (B)”. The keyword formula created by the keyword formula integrating means 43 is input to the text search formula input means 12 and used by the text search means 11 as a keyword formula for text search.

[Fourth Invention] (Explanation of Structure) FIG. 8 is a diagram showing a structure of a fourth invention of a text search apparatus according to an embodiment of the present invention. In the configuration of FIG. 8, a combination condition selection unit 27, a keyword combination unit 28, and a keyword attribute definition unit 29 are added to the configuration of FIG. 1 as new components. Like the third invention,
The combining condition selecting means 27 is not an essential constituent means.

The keyword attribute definition unit 29 is a unit for defining an attribute used as a keyword for text search or a condition for combining attributes from the DB search formula input by the DB search condition input unit 21.

The keyword combining unit 28 extracts a keyword from the DB search formula input by the DB search condition input unit 21 based on the definition of the keyword attribute defining unit 29.
This is a means for creating a keyword expression by combining them.

The keyword formula integrating means 43 combines the keyword formula created by the keyword combining means 28 with the keyword formula based on the developed word obtained by the developed word extracting means 23 under the combining condition designated by the user by the combining condition selecting means 27. I do.

(Explanation of Operation) A concrete example of the operation of the text search apparatus shown in FIG.
Will be referred to as an embodiment. ). In the DB search condition input means 21, “(maker =“ Company A ”OR
“Company B”) AND (shape = “sedan”) AND (purpose =
When a DB search formula of “family”) AND (price ≦ 2 million yen) is input, “AAAY”, “BBB” is used by the expanded word extracting means 23 in the same flow as in the first embodiment.
X ”and“ bbbbx01. ”In the keyword expression integrating unit 43, first,“ AAAY ”OR“ BBBX ”OR
The keyword formula (C) of "bbbbx01" is obtained.

The keyword attribute definition means 29 defines an attribute to be used as a keyword for text search and a connection relationship between the attributes from the attribute value input as a DB search expression by the DB search condition input means 21. For example, it is defined that an attribute having a character type such as “maker”, “shape”, and “purpose” is used as a keyword for text search. Alternatively, even if the attribute value is of a character type, a definition that the "purpose" is not a keyword may be considered. In addition, as a connection relationship between these attributes, “manufacturer” AND
(“Shape” OR “purpose”). Based on this definition, the keyword combining means 28 determines “(“ Company A ”OR“ Company B ”) AND (“ Sedan ”OR
“Family”) ”is obtained as the keyword formula (D).

The combining condition selecting means 27 selects whether the keyword formula (C) and the keyword formula (D) are combined by an OR condition or an AND condition by a user's instruction. Assuming that the OR condition is selected, the keyword expression integrating means 43 selects "((" Company A "OR" Company B ")) AND
(“Sedan” OR “Family”)) OR (“AAA
Y "OR" BBBX "OR" bbbbx01 ")" is created as a keyword expression. Keyword type integration means 43
The keyword formula created in step 1 is a text search formula input unit 1.
2 is used as a keyword expression for text search in the text search means 11.

[Fifth Invention] (Description of Structure) FIG. 9 is a diagram showing a structure of a text search apparatus according to a fifth embodiment of the present invention. Attribute development relation definition means 30 is added to the configuration of FIG. 1 as a new component. It should be noted that the present invention can be easily realized with the configuration shown in FIG. 3, FIG. 7, or FIG.

The attribute expansion relation definition means 30 is a means for defining an expansion relation between attributes. For example, an expansion relationship such as expansion of an upper word into a lower word is conceivable. The DB search statement creation unit 22 expands the DB search expression input by the DB search condition input unit 21 based on the definition of the attribute expansion relationship definition unit 30, and sets the attribute value defined by the acquisition attribute definition unit 25. Is created as a development word.

(Explanation of Operation) Regarding the operation of the text search apparatus shown in FIG. 9, it is assumed that "automobile DB" is specified in the DB search condition input means 21 and a DB search expression of "objective =" sports "" is input. An embodiment using a specific example (hereinafter, this embodiment is referred to as a fifth embodiment) will be described.

In the attribute expansion relation definition means 30,
By defining expansion relation rules such as "" Sports "(vehicle DB. Purpose) → open (vehicle DB. Shape)", D created by the DB search statement creation means 22 is defined.
“Shape =“ open ”” is added as a DB search expression to the B search expression “objective =“ sports ”” as a DB search expression under an OR condition. In other words, FIG.
The WHERE condition of (A) is “(Automobile DB. Purpose =
“Sports”) OR (automobile DB. Shape = “open”) ”. If the acquisition attribute is defined as “vehicle name” by the acquisition attribute definition means 25, the SELECT shown in FIG.
The T condition is only “SELECT: automobile DB. Car name”. The expansion word extracting means 23 executes this DB search command sentence, obtains the car names “AAX” and “BBBY” as expansion words, and the keyword expression creating means 24 outputs “AAAAX” OR
A keyword expression “BBBY” is created. That is, "B
BBY "is a keyword added according to the present invention.
The keyword formula created by the keyword formula creating unit 24 is input to the text search formula input unit 12 and used by the text search unit 11 as a keyword formula for text search.

[Sixth Invention] (Explanation of Structure) FIG. 10 is a diagram showing a structure of a text search apparatus according to a sixth embodiment of the present invention. An ambiguous word removing unit 36, an ambiguous determination rule defining unit 37, and a word dictionary 260 are added to the configuration of FIG. 1 as new components. It should be noted that the present invention can be easily realized with the configuration shown in FIG. 3, FIG. 7, or FIG.

If the expanded word matches a word existing in the word dictionary 260 or is a short word, the expanded word is considered to have ambiguity with other words. Ambiguous words are likely to have an adverse effect such as over-searching if they are added as expanded words as they are under the OR condition, so it is desirable to exclude them from keyword expressions. The ambiguity determination rule defining unit 37 is a unit for defining a condition that the expanded word obtained by the expanded word extracting unit 23 is an ambiguous word. For example, “when the word matches the word existing in the word dictionary 260” or “when the character is shorter than n characters”
Define rules such as considering them as ambiguous words and excluding them from expanded words. The ambiguous word removing means 36 is means for removing an ambiguous word from the result of the expanded word extracting means 23 according to the definition of the ambiguous determination rule defining means 37.

(Explanation of Operation) The operation of the text search apparatus shown in FIG.
Example using a specific example of a case where words “Y” and “BBBX” and model names “bbbx01” and “bb” are acquired as expansion words and a word having the same notation as the car name “AAAY” exists in the word dictionary 260. (Hereinafter, this embodiment will be referred to as a sixth embodiment.)

If the ambiguity determination rule definition means 37 defines the criterion of an ambiguity word as “word that matches a word existing in the word dictionary 260”, the ambiguity word removal means 36 Word dictionary 2 using the obtained expanded words as keys
Search for 60. Since “AAAY” is registered in the word dictionary 260, the ambiguous word removing unit 36 determines “AAAY” as an ambiguous word and excludes it from the expanded word list obtained by the expanded word extracting unit 23. Also, the ambiguity determination rule definition means 3
If a short character string of 2 characters or less is defined as a criterion for determining an ambiguous word in step 7, “bb” is excluded from the expanded word list obtained by the expanded word extraction unit 23 as an ambiguous word.
Thus, the ambiguous word removing means 36 excludes the car name “AAAY” and the type name “bb” from the expanded words, and the keyword expression creating means 24 uses ““ BBBX ”O as the keyword expression.
R “bbbx01” ”is obtained. Keyword expression creation means 2
The keyword formula created in step 4 is input to the text search formula input unit 12 and used as a keyword formula for text search in the text search unit 11.

Here, the word dictionary 260 stores at least heading (notation) information and part of speech information. Rather than simply determining an ambiguous word only based on whether or not it is registered in the word dictionary 260, the accuracy of ambiguous word determination may be further improved by using detailed part of speech information or meaning information. For example, “AAAY” is the word dictionary 26
If 0 is registered as a proper noun / vehicle name, variations such as not determining it as an ambiguous word may be considered.

[Seventh Invention] (Explanation of Structure) FIG. 11 is a diagram showing a structure of a text search apparatus according to a seventh embodiment of the present invention.
An ambiguous word determination unit 38 is added to the configuration of FIG. 10 instead of the ambiguous word removal unit 36, and a related word acquisition unit 39 and a related word definition unit 40 are added as new components. It should be noted that the present invention can be easily realized with the configuration shown in FIG. 3, FIG. 7, or FIG.

The ambiguous word judging means 38 includes the expanded word extracting means 2
This is a means for determining an ambiguous word from the result of 3 in accordance with the definition of the ambiguous judgment rule definition means 37. Ambiguous word determination means 38
The word determined not to be ambiguous is input to the keyword expression creating means 24. The word determined to be ambiguous by the ambiguous word judging means 38 is input to the related word acquiring means 39.

The related word definition means 40 is a
The attribute which becomes a related word is defined from the line including the expanded word determined to be ambiguous in FIG. The related word acquisition unit 39 acquires a related word for an ambiguous word from the DB storage unit 20 according to the definition of the related word definition unit 40, creates a keyword expression combining the ambiguous word and the related word, and creates a keyword expression To enter.

(Explanation of Operation) The operation of the text search apparatus shown in FIG.
Y ("BBBX", model names "bbbb01" and "bb" are acquired as expansion words, and an embodiment using a specific example of a case where a word having the same notation as the car name "AAAY" exists in the word dictionary 260 ( Hereinafter, this embodiment is referred to as a seventh embodiment.)

The ambiguous word judging means 38 calculates "AAAY" (car DB. Car name) in the same manner as in the sixth embodiment.
“Bb” (automobile DB. Model name) is determined to be an ambiguous word.

The related word definition means 40 defines related words such as "automobile DB. Maker", "PC DB. Maker", and "PC DB. Software name" for the ambiguous words obtained by the ambiguous word judging means 38. Keep it. The ambiguous word determining means 38 determines that “AAAY” and “bb”
, And the related word definition means 40 defines that the ambiguous word of the automobile DB is to use the value of “automobile DB. Maker” as a related word. Then, the related word acquiring means 39 obtains “company A” as a related word, and converts the ambiguous words “AAAY” and “bb” into respective related words “A”.
Company "by an AND combination, and the keyword expression"("AAAY" AND "company A") OR ("bb" AN
D “Company A”) ”and input it to the keyword formula creating means 24. Here, not only an AND connection but also a keyword expression by proximity operation of a NEAR connection can be considered.

The keyword expression creating means 24 OR-joins the keyword expression "" BBBX "OR" bbbx01 "" obtained by ORing the expanded words inputted from the expanded word extracting means 23 with the keyword expression created by the related word acquiring means 39. To create a keyword expression. That is, as a keyword expression,
“(“ BBBX ”OR“ bbbbx01 ”) OR (“ AA
AY ”AND“ Company A ”) OR (“ bb ”AND“ A
Company ")". The keyword formula created by the keyword formula creating unit 24 is input to the text search formula input unit 12 and used by the text search unit 11 as a keyword formula for text search.

[Eighth Invention] (Description of Structure) FIG. 12 is a diagram showing a structure of an eighth invention of the text search apparatus according to the embodiment of the present invention.
A development result editing unit 41 is added to the configuration of FIG. 7 as a new component. The present invention can be easily realized by the configuration shown in FIG. 1, FIG. 3, or FIG.

The expansion result editing means 41 is a means for displaying all keyword expressions to be input including expansion words, and providing an interface for the user to edit such as addition or deletion.

(Explanation of Operation) The operation of the text search apparatus shown in FIG. 12 will be described with reference to a specific example in an embodiment (hereinafter, this embodiment is referred to as an eighth embodiment). Third
FIG. 26 shows a display example of the development result of the embodiment of the present invention. In the upper part of the screen of FIG. 26, the keyword expression ““ Company A ”AND“ Sedan ”” input by the text search condition input means 26 is displayed.
Is displayed. The keyword can be corrected, added, or deleted from the upper part of the screen. At the bottom of the screen in FIG. 26,
A list of expansion words is displayed. Each expansion word is, for example, OR
Are joined by operators. For example, type name "aaax01"
If you want to delete “aaax02”, use a keyboard or a mouse or other input means to delete “aaaax01” or “aaaax02”.
Select and delete to delete. "O" at the center of the screen
The box of "R" indicates the keyword expression input by the text search condition input means 26 and the keyword expression "AAA"
X "OR" and "AAAY"".
The user presses an inverted triangle (▽) button on the right to display “A
It is also possible to select the operator "ND". By editing the keyword expression and pressing the execute button in this way,
The expansion result editing means 41 uses the keyword expression “(“ Company A ”AN
D “Sedan”) OR (“AAX” OR “AAA”
Y ")" is created and input to the text search formula input means 12. The text search means 11 uses this keyword formula for text search.

[Ninth Invention] (Explanation of Structure) FIG. 13 is a diagram showing a structure of a ninth invention of a text search device according to an embodiment of the present invention.

The text search condition input means 26 is a means for a user to input a keyword for text search.

The similarity determination definition means 32 is a means for defining similarity conditions for determining that rows in a database are similar to each other based on the value of an attribute.

The similar keyword acquiring unit 31 searches for a line including the keyword input by the text search condition input unit 26, and finds a line where the value of the attribute name defined by the similarity definition unit 32 satisfies the similar condition. From the attribute having the keyword input by the user as a value.

The keyword formula integrating means 43 creates a keyword formula obtained by integrating the keyword formula obtained by combining the keywords input to the text search condition input means 26 and the keyword formula obtained by the similar keyword obtaining means. Means.

(Explanation of Operation) The operation of the text search apparatus shown in FIG. 13 will be described with reference to an example in which a keyword "AAX" is input in the text search condition input means 26 (hereinafter, this embodiment will be described in the following). Ninth Embodiment) will be described.

The similar keyword acquiring means 31 inputs the keyword “AAA” inputted by the text search condition inputting means 26.
“X” as a key to “AAA” from the DB in the DB storage means 20
A line containing X "is searched, and a keyword is extracted from a line similar to the line.

The criterion for the similarity determination is the similarity determination definition means 32
Defined by The similarity determination definition means 32 defines, for example, attributes of “maker” and “shape” as attributes used for the criteria of similarity determination, and regards rows having values that match all of these attributes as similar rows. Define similar conditions. As similarity conditions, besides "all match", the attribute values defined as the similarity criterion may be "match more than a certain number". For example, as the attributes of the similarity determination criterion, “manufacturer”, “shape”, “purpose”, and “engine name” are defined, and the similarity condition of “match three or more” is defined. "Engine name"
Are determined to be similar even if the values of “purpose” do not match. Alternatively, a similar condition such as "the values are close" may be considered instead of the similar condition "the values match". If you define the attribute "price" as a similarity criterion,
A range such as “price ≦ 2,000,000 yen” may be used as the similar condition.

From FIG. 15, the row containing “AAX” is “Maker =“ Company A ””, “Shape =“ Sedan ”” or “Maker =“ Company A ””, and “Shape =“ Coupe ””. The rows that match these are the rows of “vehicle name =“ AAAY ”” and “vehicle name =“ AAAZ ””. That is, the similar keyword acquiring unit 31 sets the “AAAX” line and the “AA
It is determined that the rows of “AY” and “AAAZ” are similar, and “AA” is determined from the rows of “AAAY” and “AAAZ”.
"AAAY" and "AA" with the same attribute (car name) as "AX"
AZ "is obtained as a keyword.

The keyword expression integrating means 43 inputs the keyword “AAA” input to the text search condition inputting means 26.
X and the keywords “AAAY” and “AAAZ” acquired by the similar keyword acquisition means 31 are combined, for example, “AAAAX” OR “AAAY” OR “AAA”
A keyword expression such as "Z""is created. The keyword formula created by the keyword formula integrating means 43 is input to the text search formula input means 12 and used by the text search means 11 as a keyword formula for text search.

The first to ninth inventions of the present invention have been described above, but each invention can be realized by a computer program. For example, in the case of the first invention, it is necessary for the computer to make each of the DB search condition input means 21, the DB search command sentence creation means 22, the expansion word extraction means 23, the keyword expression creation means 24 and the acquisition attribute definition means 25 function. A form provided by a computer-readable recording medium on which the program is recorded may be used. In addition, the second
The same applies to the ninth invention.

[0120]

According to the text search apparatus of the present invention, which is characterized by keyword expansion of text search using a tabular database (for example, RDB), it is possible to expand keywords into specific names and expand in consideration of numerical conditions. It is possible to realize a high-precision text search with little omission for various search requests.

Further, since a plurality of classification viewpoints can be managed by a single table, the maintenance viewpoint is superior to a thesaurus having a tree structure. Furthermore, RDBs are widely used in databases such as product information, so they are highly diversified. In particular, specific names (such as product names) and numerical information (such as product specifications) are easier to obtain than a thesaurus. Are better.

[Brief description of the drawings]

FIG. 1 is a diagram illustrating a configuration of a text search apparatus according to a first embodiment of the present invention;

FIG. 2 is a diagram showing an example of a specific implementation for inputting a DB name and a DB search formula by a DB search condition input unit 21 in FIG. 1;

FIG. 3 is a diagram showing a configuration of a second invention of a text search device according to an embodiment of the present invention.

4 is a diagram showing an example of a specific implementation of a DB search condition input unit 21 in FIG.

5 is a diagram showing another example of a specific implementation of the DB search condition input unit 21 in FIG.

FIG. 6 is a diagram showing an example of an implementation example of a text search condition input means 26 for enabling input of a natural language.

FIG. 7 is a diagram showing a configuration of a third invention of the text search device in the embodiment of the present invention.

FIG. 8 is a diagram showing a configuration of a fourth invention of the text search device in the embodiment of the present invention.

FIG. 9 is a diagram illustrating a configuration of a text search apparatus according to a fifth embodiment of the present invention;

FIG. 10 is a diagram illustrating a configuration of a text search apparatus according to a sixth embodiment of the present invention;

FIG. 11 is a diagram showing a configuration of a text search apparatus according to a seventh embodiment of the present invention;

FIG. 12 is a diagram showing a configuration of an eighth invention of the text search device in the embodiment of the present invention.

FIG. 13 is a diagram showing a configuration of a ninth invention of a text search device according to an embodiment of the present invention.

FIG. 14 is a flowchart showing a flow of automatic creation of a menu in the DB search expression input means 211.

FIG. 15 is a diagram illustrating an example of an automobile DB.

FIG. 16 shows a D created by a DB search statement creation unit 22;
7 shows an example of a format of a B search command sentence and a specific example thereof.

FIG. 17 is a diagram showing an example of a list of DBs displayed by a DB user selection unit 210.

18 is a diagram showing an example of an attribute value input menu created by the DB search expression input means 211 of FIG.

FIG. 19 is a diagram showing an example of an attribute value input menu created by the DB search expression input means 211 of FIG. 2;

20 is a diagram illustrating an example of an attribute value input menu created by the DB search expression input unit 211 in FIG. 2;

21 is a diagram showing an example of an attribute value input menu created by the DB search expression input means 211 of FIG.

FIG. 22 is a diagram illustrating an example of HTML text that implements the menu of FIG. 17;

FIG. 23 is a diagram showing an example of the contents of a controlled word dictionary.

FIG. 24 is a diagram showing an example of contents of a developed word table 213.

FIG. 25 is a diagram showing an example of contents of a user information memory 214.

FIG. 26 is a diagram illustrating a display example of a development result of the third embodiment.

[Description of sign]

 DESCRIPTION OF SYMBOLS 10 Text storage means 11 Text search means 12 Text search formula input means 13 Search result display means 20 DB storage means 21 DB search condition input means 22 DB search command sentence creation means 23 Expansion word extraction means 24 Keyword formula creation means 25 Acquisition attribute definition Means 26 Text search condition input means 27 Join condition search means 28 Keyword integration means 29 Keyword attribute definition means 30 Attribute expansion relation definition means 36 Ambiguity word removal means 37 Ambiguity determination rule definition means 38 Ambiguity word determination means 39 Related word acquisition means 40 Related Word definition means 41 Expansion result editing means 42 DB search condition creation means 43 Keyword expression integration means 210 DB user selection means 211 DB search expression input means 212 DB automatic selection means 213 Developed word table 215 DB search expression creation means 260 Word dictionary 261 Search request sentence input means 262 Sentence analysis means 263 Control processing means 264 Controlled word dictionary 265 Numeric expression extraction means 266 Numeric extraction rule definition means 267 Numeric condition extraction means 268 Numeric conversion rule definition means

 ──────────────────────────────────────────────────続 き Continuing from the front page (72) Inventor Shunichi Fukushima 5-7-1 Shiba, Minato-ku, Tokyo F-term in NEC Corporation (Reference) 5B075 ND03 ND40 NK02 NK46 NK54 PP13 PP22 PP24 PP30 PQ02 PQ20 PQ40 PQ63 PR03 PR06 QM08 QP03 QP10 UU08 UU40

Claims (30)

[Claims] 1. A DB search condition input means for designating a tabular database consisting of rows and columns and inputting a database search formula for searching a text database, and a keyword to be acquired in the tabular database as a value An acquisition attribute defining means for defining a column item to have, and selecting a row satisfying the database search formula from a tabular database specified by the DB search condition input means,
DB search statement creation means for creating a database search statement for acquiring a keyword which is a value of a column item defined by the acquisition attribute definition means from the row; and executing the database search statement to obtain the acquisition attribute. Expansion word extraction means for acquiring a keyword that is a value of a column item defined by the definition means; and keyword expression creation means for creating a keyword expression for searching the text database using the keyword acquired by the expansion word extraction means. A text search device having at least a text search device. 2. The DB search condition input means displays a list of the tabular databases, and provides a user interface for selecting the tabular databases. The DB user selecting means selects the tabular databases. And a DB search formula input means for providing a user interface for acquiring a column item from the tabulated database, displaying the list, and inputting a value for each column item. Item 1. The text search device according to Item 1. 3. A text search condition input means for inputting a keyword or a natural language sentence for a text search by a user, and an appropriate keyword or natural language sentence input through the text search condition input means. DB search condition creating means for selecting a tabular database composed of rows and columns and creating a database search expression that is an expression describing a condition for searching for a desired row from the tabular database; An acquisition attribute definition unit that defines a column item having a keyword to be acquired as a value, and a row that satisfies the database search formula is selected from a tabular database selected by the DB search condition creation unit.
DB search statement creation means for creating a database search statement for acquiring a keyword which is a value of a column item defined by the acquisition attribute definition means from the row; and executing the database search statement to obtain the acquisition attribute. Expansion keyword extraction means for acquiring a keyword of a column item defined by the definition means; and keyword expression creation means for creating a keyword expression for text search using the keyword acquired by the expansion word extraction means. Characteristic text search device. 4. The system according to claim 1, wherein the DB search condition creating means includes: a developed word table for storing information indicating a tabular database to which the keyword belongs and column items of the keyword in the tabular database; Automatic DB selecting means for selecting the tabular database by collating the keyword input from the user or a keyword obtained from a natural language sentence input by the text search condition input means with the developed word table; 4. A database search formula creating means for creating a database search formula from a keyword used for database selection by an automatic selection means and a column item of the keyword in a tabular database. Text search device. 5. A user information memory for storing personal information such as usage history of the text search device and user interest information,
5. The text search apparatus according to claim 4, further comprising: selecting the tabular database with reference to the user information memory. 6. A text search condition input means for inputting a keyword or a natural language sentence for text search by a user, and an appropriate keyword or natural language sentence input through the text search condition input means. DB search condition creating means for selecting a tabular database composed of rows and columns and creating a database search expression that is an expression describing a condition for searching for a desired row from the tabular database; An acquisition attribute definition unit that defines a column item having a keyword to be acquired as a value, and a row that satisfies the database search formula is selected from the tabular database selected by the DB search condition creation unit, and the acquisition attribute definition is selected from the row. Create a database search statement to retrieve the keyword that is the value of the column item defined by DB search command sentence creating means for executing, the expanded word extracting means for executing the database search command sentence to obtain a keyword which is a value of a column item defined by the acquisition attribute defining means, and inputting by the text search condition input means A keyword expression combining the extracted keywords or a keyword expression combining keywords obtained from the natural language sentence input by the text search condition input means and a keyword expression combining the keywords obtained by the expansion word extraction means are integrated. And a keyword expression integrating means for creating a keyword expression for text search. 7. A keyword attribute defining means for defining a column item or a joining condition between column items in a tabular database having, as a value, a keyword used for a text search from a database search formula input via said DB search condition input means. And a keyword combining unit that creates a keyword formula by extracting and combining keywords from the database search formula based on the definition of the keyword attribute defining unit. The keyword formula creating unit includes: Integrating the obtained keyword expression and a keyword expression obtained by combining the keywords obtained by the expanded word extraction means,
2. The text search apparatus according to claim 1, wherein the text search apparatus is a keyword formula integration unit that creates a keyword formula for searching the text database. 8. An attribute expansion relation defining means for defining an expansion relation between column items in a tabular database, wherein said DB search command sentence creating means is a database inputted by said DB search condition input means. Search formula or D
B expands the database search formula created by the search condition creating means based on the definition of the attribute expansion relationship defining means,
7. The text search device according to claim 1, wherein a database search command sentence for obtaining a keyword that is a value of a column item defined by the obtainment attribute defining unit is created. 9. An ambiguity determination rule defining means for defining a criterion for determining whether or not a keyword acquired by said expanded word extraction means is an ambiguous word, and an ambiguity determination rule definition means based on the keyword acquired by said expanded word extraction means. 4. The text according to claim 1, further comprising: an ambiguous word removing unit that removes an ambiguous word based on a defined criterion and outputs the keyword from which the ambiguous word has been removed to the keyword expression creating unit. Search device. 10. An ambiguity determination rule defining means for defining a criterion for determining whether a keyword obtained by said expanded word extraction means is an ambiguous word, and said ambiguity determination rule definition means based on the keyword obtained by said expansion word extraction means. 7. The text search apparatus according to claim 6, further comprising: an ambiguous word removing unit that removes an ambiguous word based on a defined criterion and outputs the keyword from which the ambiguous word has been removed to the keyword formula integrating unit. 11. An ambiguity determination rule defining means for defining a criterion for determining whether a keyword acquired by said expanded word extraction means is an ambiguous word, and an ambiguity determination rule definition means based on the keyword acquired by said expanded word extraction means. An ambiguous word judging unit for judging an ambiguous word based on the defined judging criterion, and an extraction condition of a value of a column item which is a related word for the ambiguous word is defined from a line obtained by the expanded word extracting unit for the ambiguous word. Related word definition means, and related word acquisition means for acquiring a related word for the ambiguous word according to the definition of the related word definition means, and creating a keyword formula by combining the ambiguous word and the related word. Wherein the keyword expression creating means creates a keyword expression for searching the text database, also using the keyword expression created by the related word acquiring means. The text search device according to claim 1 or 3, wherein 12. An ambiguity determination rule defining means for defining a criterion for determining whether a keyword obtained by said expanded word extraction means is an ambiguous word, and said ambiguity determination rule definition means based on the keyword obtained by said expansion word extraction means. An ambiguous word judging unit for judging an ambiguous word based on the defined judging criterion, and an extraction condition of a value of a column item which is a related word for the ambiguous word is defined from a line obtained by the expanded word extracting unit for the ambiguous word. Related word definition means, and related word acquisition means for acquiring a related word for the ambiguous word according to the definition of the related word definition means, and creating a keyword formula by combining the ambiguous word and the related word. 7. The method according to claim 6, wherein the keyword formula integrating means also integrates the keyword formula created by the related word acquiring means to create a keyword formula for text search. Text search device. 13. An expansion result editing means for displaying all the keyword expressions created by said keyword expression creating means and providing an interface for a user to edit such as addition / deletion. 3. The text search device according to 3. 14. A system according to claim 6, further comprising an expansion result editing means for displaying all the keyword expressions created by said keyword expression integrating means and providing an interface for a user to perform an addition or deletion. Text search device. 15. A text search condition input means for inputting a keyword for text search by a user, and in a tabular database composed of rows and columns, the similarity between rows is determined based on the values of column items in the tabular database. A similarity determination defining means for defining a similar condition for performing the search, a line including the keyword input by the text search condition input means is searched from the tabular database, and from the line and the line satisfying the similarity condition, A similar keyword obtaining unit that obtains, as a keyword, a value of a column item having the input keyword as a value; a keyword formula that combines the keywords input through the text search condition input unit; and a similar keyword obtaining unit that obtains the keyword expression. Create a keyword expression that combines the keyword expression that combines the keywords Text search apparatus characterized by having at least a keyword expression integration unit. 16. A DB search condition input function which allows a computer to specify a tabular database composed of rows and columns and to input a database search formula for searching a text database, and a keyword to be acquired in the tabular database An acquisition attribute definition function for defining a column item having a value as a value, and selecting a row that satisfies the database search expression from a tabular database specified by the DB search condition input function,
A DB search statement creation function for creating a database search statement for acquiring a keyword that is a value of a column item defined by the acquisition attribute definition function from the row; and executing the database search statement to obtain the acquisition attribute. An expansion word extraction function for acquiring a keyword that is a value of a column item defined by a definition function; and a keyword expression creation function for creating a keyword expression for searching the text database using the keyword acquired by the expansion word extraction function. A computer-readable recording medium that records at least a text search program to be realized. 17. The DB search condition input function displays a list of the tabular databases, and provides a user interface for selecting the tabular database, and a DB user selecting function to select the tabular database. 17. The text according to claim 16, further comprising: a DB search expression input function for acquiring a column item from the tabulated database and displaying the list, and providing a user interface for inputting a value for each column item. A computer-readable recording medium recording a search program. 18. A computer, comprising: a text search condition input function for inputting a keyword or a natural language sentence for text search by a user; and a keyword or a natural language sentence input through the text search condition input function. A DB search condition creation function for selecting a tabular database including appropriate rows and columns, and creating a database search expression that is an expression describing a condition for searching for a desired row from the tabular database; An acquisition attribute definition function that defines a column item having a keyword to be acquired in a format database as a value, and a row that satisfies the database search formula is selected from a table database selected by the DB search condition creation function
A DB search statement creation function for creating a database search statement for acquiring a keyword that is a value of a column item defined by the acquisition attribute definition function from the row; and executing the database search statement to obtain the acquisition attribute. In order to realize at least a development word extraction function of acquiring a keyword of a column item defined by a definition function, and a keyword expression creation function of creating a keyword expression for text search using the keyword acquired by the expansion word extraction function And a computer-readable recording medium recording the text search program. 19. A function for creating a DB expansion condition table for storing information indicating a tabular database to which a keyword belongs and a column item of the keyword in a tabular database in the tabular database; A DB automatic selection function of selecting the tabular database by comparing the keyword input from the input function or the keyword obtained from the natural language sentence input by the text search condition input function with the developed word table 19. The database search function creating function for creating a database search equation from a keyword used for database selection by the DB automatic selection function and a column item of the keyword in a tabular database in a tabular database. Computer readable recorded text search program Recording medium. 20. The apparatus further comprising a function of creating a user information memory for storing personal information such as interest information of a user's usage history, and wherein the DB automatic selection function refers to the user information memory and stores the tabular database. 20. A computer-readable recording medium on which the text search program according to claim 19 is selected. 21. A text search condition input function for inputting a keyword or a natural language sentence for text search by a user, and an appropriate keyword or natural language sentence input through the text search condition input function. A DB search condition creation function for selecting a tabular database composed of rows and columns and creating a database search expression that is an expression describing a condition for searching for a desired row from the tabular database; An acquisition attribute definition function that defines a column item having a keyword to be acquired as a value, and a row that satisfies the database search formula is selected from the tabular database selected by the DB search condition creation function, and the acquisition attribute definition is selected from the row. A database search statement to retrieve the keyword that is the value of the column item defined in the function A DB search statement creation function to be executed, an expansion word extraction function for executing the database search statement and acquiring a keyword that is a value of a column item defined by the acquisition attribute definition function, and a text search condition input function. A keyword expression combining the input keywords or a keyword expression combining keywords obtained from the natural language sentence input by the text search condition input function and a keyword expression combining the keywords obtained by the expansion word extraction function A computer-readable recording medium that records a text search program for realizing at least a keyword expression integration function of integrating and creating a keyword expression for text search. 22. A computer that defines a column item or a join condition between column items in a tabular database having a keyword used for text search as a value from a database search formula input via the DB search condition input function. An attribute definition function; and a keyword combining function of creating a keyword expression by extracting and combining keywords from the database search expression based on the definition of the keyword attribute definition function. Integrate the keyword expression obtained by the combining function and the keyword expression obtained by combining the keywords obtained by the expansion word extraction function,
17. The computer-readable recording medium storing a text search program according to claim 16, which is a keyword expression integration function for creating a keyword expression for searching the text database. 23. A computer, further comprising: an attribute expansion relation definition function for defining an expansion relation between column items in a tabular database, wherein the DB search command statement creation function is input by the DB search condition input function. Database search formula or D
B Expand the database search formula created by the search condition creation function based on the definition of the attribute expansion relationship definition function,
22. A computer-readable recording program for recording a text search program according to claim 16, wherein a database search command sentence for obtaining a keyword which is a value of a column item defined by the acquisition attribute definition function is created. Possible recording medium. 24. A computer, comprising: an ambiguity determination rule definition function for defining a criterion for determining whether a keyword obtained by the expansion word extraction function is an ambiguous word; and an ambiguity determination rule from the keyword obtained by the expansion word extraction function. 19. The text according to claim 16, further comprising: an ambiguous word removing function of removing an ambiguous word based on a criterion defined by a definition function and outputting the keyword from which the ambiguous word has been removed to the keyword expression creating function. A computer-readable recording medium recording a search program. 25. An ambiguity determination rule defining means for defining a criterion for determining whether or not a keyword acquired by the expansion word extraction function is an ambiguous word; and an ambiguity determination rule based on the keyword acquired by the expansion word extraction function. 22. The text search according to claim 21, further comprising: an ambiguous word removing function that removes an ambiguous word based on a criterion defined by a definition function and outputs the keyword from which the ambiguous word has been removed to the keyword expression integration function. A computer-readable recording medium on which a program is recorded. 26. A computer, comprising: an ambiguity determination rule definition function for defining a criterion for determining whether a keyword acquired by the expansion word extraction function is an ambiguous word; and an ambiguity determination rule from the keyword acquired by the expansion word extraction function. An ambiguous word judging function for judging an ambiguous word based on a judgment criterion defined by a definition function; and an extraction condition of a value of a column item serving as a related word for the ambiguous word from a row obtained by the expanded word extracting function for the ambiguous word. A related word definition function for defining, and a related word acquisition function for acquiring a related word for the ambiguous word according to the definition of the related word definition function, and creating a keyword formula by combining the ambiguous word and its related word; A keyword for searching the text database by using the keyword expression created by the related word acquisition function. A computer-readable recording medium according to claim 16 or claim 18, wherein the text retrieval program for creating a. 27. An ambiguity determination rule definition function for defining a criterion for determining whether a keyword acquired by the expansion word extraction function is an ambiguous word, and an ambiguity determination rule definition function based on the keyword acquired by the expansion word extraction function. An ambiguous word determination function for determining an ambiguous word based on the defined criteria, and an extraction condition for a value of a column item serving as a related word for the ambiguous word from a line obtained by the expanded word extraction function for the ambiguous word. Further realizing a related word definition function and a related word obtaining function of obtaining a related word for the vague word according to the definition of the related word definition function, and creating a keyword formula by combining the vague word and the related word. 22. The text according to claim 21, wherein the keyword expression integration function integrates the keyword expression created by the related word acquisition function to create a keyword expression for text search. A computer-readable recording medium on which a program for retrieving a list is recorded. 28. The computer system according to claim 16, further comprising: displaying, on the computer, all of the keyword expressions created by the keyword expression creating function, and providing an expansion result editing function of providing an interface for a user to perform editing such as addition or deletion. 19. A computer-readable recording medium recording the text search program according to 18. 29. The computer system according to claim 21, further comprising: displaying, on a computer, all of the keyword expressions created by the keyword expression integration function, and providing an expansion result editing function of providing an interface for a user to perform an addition / deletion operation. And a computer-readable recording medium recording the text search program. 30. A computer, comprising: a text search condition input function for inputting a keyword for text search by a user; and, in a tabular database comprising rows and columns, similarity between rows by a column item value of the tabular database. A similarity determination definition function for defining similarity conditions for determining the similarity, and a line including the keyword input by the text search condition input function is searched from the tabular database, and the line and the line satisfying the similarity condition are searched. A similar keyword acquisition function for acquiring a value of a column item having a keyword input by a user as a value as a keyword, a keyword expression combining keywords input via the text search condition input function, and a similar keyword acquisition function Key that combines the keyword expression obtained by combining with the keyword expression A computer-readable recording medium on which a text search program for realizing at least a keyword expression integration function for creating a word expression is recorded.