JPH09190451A - Information retrieving device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain a device whereby a retrieval object which is not intended by a retrieval expression by a retriever is not outputted. SOLUTION: A database 103 stores the retrieval objects. A retrieval executing part 102 retrieves the retrieval objects of the database 103 based on the retrieval expression inputted from an input part 101. A retrieval object comparing part 104 gives a priority degree to the retrieval objects which are added in a set expressed by the retrieval expression through the use of an association value to the set of additional keywords which are added to the retrieval object from the set of the input keywords which are added to the retrieval expression so as to permit the retrieval object more intended by the retriever to have a higher numerical value. An output part 105 outputs the retrieval object to in accordance with the priority degree which is given to the retrieval object outputted by the retrieval executing part 102 in the retrieval object comparing part 104.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an information retrieval apparatus for retrieving an information retrieval target such as a document from a retrieval formula.

[0002]

2. Description of the Related Art Conventionally, in a device of this type, as shown in, for example, Japanese Patent Laid-Open No. 6-259480, until the number of search objects included in a set represented by a search expression exceeds a certain threshold value. The search formula is changed so that the search range is widened, and when the search formula is equal to or larger than a certain threshold, all the search targets included in the set represented by the search formula at that time are output.

[0003]

By the way, the database stores many search objects. Some additional keywords that symbolize the information are attached to the search target, and depending on what those additional keywords are,
The search target included in the set represented by the search expression is determined. Further, the searcher determines the search formula without knowing which additional keyword is attached to each search target.

As described above, in the conventional technique, all search objects included in the set represented by the search expression are output. Then, the searcher determines the search formula without knowing which additional keyword is attached to each search target,
There was a problem that the searcher outputs an unintended search target in the search formula. In addition, when the number of search targets included in the set expressed by the search formula exceeds a certain threshold, even if the set includes search targets that the searcher did not intend in the search formula, There was a problem that other search targets intended by the search formula were not output. Further, there is a problem that the device takes a long time to respond.

From this point of view, it has been desired to realize an information retrieval device which can prevent a retrieval person from outputting an unintended retrieval target in a retrieval formula.

[0006]

The present invention employs the following structure to solve the above-mentioned problems. <Structure of claim 1> An input unit for inputting a search expression,
A search execution unit that executes a search for a search target in the database based on the database that stores the search target and the search formula that is input from the input unit, and that outputs the search target that is included in the set represented by the search formula And the associative value from the set of input keywords included in the search formula to the set of additional keywords added to the search target, the search target is included in the set expressed by the search formula. For a search target intended by a person, the search target comparison unit that gives a higher priority to the search target and the search target output from the search execution unit outputs the search target according to the priority assigned by the search target comparison unit. An output section is provided.

<Structure of Claim 2> According to the invention of Claim 2, the search target comparison unit in Claim 1 stores an association dictionary for storing an association value from a set of keywords to a set of keywords, and a set of input keywords. The associative value of the set from the keyword to the additional keyword is obtained using the associative dictionary, and the associative value is higher for the search target intended by the searcher than for the search target included in the set expressed by the search expression. It is characterized in that it is provided with a priority addition unit which is given as a numerical priority.

<Explanation of Claims 1 and 2> When a searcher inputs a search expression from the input unit, the search execution unit searches the database based on the search expression and outputs the searched search object. Further, the search target comparison unit obtains the association value for the input keyword included in the search expression and the additional keyword added to the search target searched and output by the search execution unit using the associative dictionary, Calculate priority. The output unit outputs, for example, a search target whose priority is higher than a certain threshold. Therefore, it is possible to prevent the searcher from outputting an unintended search target in the search formula.

<Structure of Claim 3> An input unit for inputting a search formula, a database for storing the search target, and a search for the search target in the database are executed based on the search formula input from the input unit. , A search execution unit that outputs the search target included in the set expressed by the search expression, and a keyword selection that selects a set of input keywords and a set of additional keywords that do not affect the results of the search in the search execution unit Section, an association dictionary that stores the association value from a certain keyword set to a certain keyword set, and the association value from the set of input keywords selected by the keyword selection unit to the set of additional keywords At the same time as obtaining, the association value is set as the priority that has a higher numerical value for the search target intended by the searcher with respect to the search target included in the set expressed by the search formula. A search target comparison unit including a priority addition unit to be attached, and an output unit that outputs the search target according to the priority assigned to the search target comparison unit with respect to the search target output from the search execution unit. It is characterized by.

<Explanation of Claim 3> When a searcher inputs a search expression from the input section, the search execution section searches the database based on the search expression and outputs the searched search object. Also, the keyword selection unit in the search target comparison unit selects a set of input keywords and a set of additional keywords that do not affect the result of the search by the search execution unit.
The priority adding unit obtains an association value from the set of input keywords selected by the keyword selecting unit to the set of additional keywords by using the associative dictionary, and the association value is set to the set expressed by the search expression. For the included search target,
The search target intended by the searcher is given a higher priority value. The output unit outputs, for example, a search target whose priority is higher than a certain threshold.

Therefore, since it is possible to give a more appropriate priority to the inventions of claims 1 and 2, it is possible to further prevent the searcher from outputting an unintended search target in the search formula. Can be large.

<Structure of Claim 4> An input unit for inputting a search formula, a database for storing a search target, and a search for the search target in the database are executed based on the search formula input from the input unit. , A search execution unit that outputs a search target included in the set expressed by the search expression, an associative dictionary that stores an association value from a keyword to a keyword, and an input keyword based on the association value of the association dictionary The priority calculation unit that calculates the associative value from the keyword to the additional keyword and calculates the priority from this associative value, and the priority calculated by the priority calculation unit for the search target included in the set expressed by the search The search target comparison unit that includes a priority addition unit that assigns a higher priority to the search target that the searcher intends, and the search target ratio that is output from the search execution unit. It is characterized in that an output unit that outputs the search target according granted the priority part.

<Explanation of Claim 4> In the invention of Claim 4, the invention of Claim 1 gives priority by using the association value from the set of input keywords to the set of additional keywords.
If there are multiple input keywords and additional keywords,
The association value is obtained for each keyword, and the priority is given based on these association values.

When a searcher inputs a search expression from the input section,
The search execution unit searches the database based on the search expression and outputs the searched search target. Also, the priority calculation unit in the search target comparison unit calculates the priority from the association value from the input keyword to the additional keyword, and the priority addition unit compares the search target included in the set expressed by the search expression. Then, based on the priority calculated by the priority calculation unit, the search target intended by the searcher is given a higher priority. The output unit outputs, for example, a search target whose priority is higher than a certain threshold.

Therefore, for the inventions of claims 1 and 2, if the associative dictionary stores the same number of associative values, a more appropriate priority can be given, so that the searcher can use the search expression. It is possible to further enhance the effect of not outputting an unintended search target.

<Structure of Claim 5> According to the invention of Claim 5, in the information retrieval apparatus according to Claim 4, the associative dictionary for storing the associative value from a certain keyword to the certain keyword is read from the associative dictionary. Based on the cache memory that stores the association value and the association value stored in the association dictionary or the association value stored in the cache memory, the association value from the input keyword to the additional keyword is obtained, and the priority is calculated from this association value. Based on the priority calculated by the priority calculation unit, the priority calculation unit and the search target included in the set expressed by the search formula are given higher priority as the search target intended by the searcher. The present invention is characterized by including a search target comparison unit including a priority addition unit to be attached.

<Explanation of Claim 5> According to the invention of Claim 5, in addition to the search object comparison unit of the invention of Claim 4, a cache memory for storing an association value is further provided. Thus, if the desired associative value is stored in the cache memory, the priority calculation unit obtains the associative value from the input keyword to the additional keyword based on the associative value, and calculates the priority from this associative value. . Therefore, each associative value search process in claim 4 can be speeded up.

<Structure of Claim 6> An input unit for inputting a search formula, a database for storing a search target, and a search for the database are executed based on the search formula input from the input unit. , A search execution unit that outputs a search target included in a set expressed by a search expression, an associative dictionary that stores an association value from a certain keyword set to a certain keyword set, and a subset of the certain set The keyword set calculation unit and the association value from the subset of the set of input keywords calculated by the keyword set calculation unit to the subset of the set of additional keywords are obtained based on the association dictionary,
Based on the priority calculated by the priority calculation unit, the priority calculation unit that calculates the priority from the association value and the search target included in the set expressed by the search formula based on the priority calculated by the priority calculation unit. Outputs the search target according to the priority given by the search target comparison unit with respect to the search target output by the search execution unit, and the search target comparison unit that has a priority addition unit that gives higher priority to the target And an output unit that operates.

<Explanation of Claim 6> In the invention of Claim 6, while the invention of Claims 4 and 5 gives priority by using the association value from each input keyword to each additional keyword, The association value is obtained for each subset of the set of additional keywords from the subset of the set of keywords, and priority is given based on these association values.

When a searcher inputs a search expression from the input section,
The search execution unit searches the database based on the search expression and outputs the searched search target. Further, the keyword set calculation unit in the search target comparison unit obtains a subset of the set of input keywords and a subset of the set of additional keywords. The priority calculation unit calculates the priority from the association value from the subset of the set of input keywords to the subset of the set of additional keywords, and the priority addition unit includes the search included in the set expressed by the search formula. Based on the priority calculated by the priority calculation unit, the search target intended by the searcher is given a higher priority. The output unit outputs, for example, a search target whose priority is higher than a certain threshold.

The associative value is expressed more as an associative value from a set of certain keywords to a set of certain keywords than by an associative value from a certain keyword to a certain keyword alone. Therefore, for the inventions of claims 4 and 5, if the associative dictionary stores the same number of associative values, a more appropriate priority can be given, so that the searcher does not intend in the search expression. It is possible to further increase the effect of not outputting the search target.

<Structure of Claim 7> An input unit for inputting a search formula, a database for storing the search target, and a search for the search target in the database are executed based on the search formula input from the input unit. , A search execution unit that outputs a search target included in a set expressed by a search expression, an associative dictionary that stores an association value from a certain keyword set to a certain keyword set, and another set different from a certain set And a keyword set calculation unit that calculates, and an association value from another set different from the set of input keywords calculated by the keyword set calculation unit to another set different from the set of additional keywords,
Based on the priority calculated by the priority calculation unit for the search target included in the set expressed by the search formula and the priority calculation unit that calculates the priority from this association value based on the association value , The search target comparison unit having a priority addition unit that gives a higher priority to the search target intended by the searcher, and the search target comparison unit added to the search target output from the search execution unit. An output unit that outputs a search target according to the priority is provided.

<Explanation of Claim 7> According to the invention of claim 7, the keyword set calculator of the invention of claim 6 calculates a subset of a certain set, while another set different from a certain set is calculated. It features a point to calculate. Therefore, even if the associative value of a subset of a certain set is not stored in the associative dictionary, the associative value can be calculated by appropriately setting another set different from that set.

<Structure of Claim 8> An input unit for inputting a search expression, a database for storing a search target, and a search for the search target in the database are executed based on the search expression input from the input unit. , A search execution unit that outputs the search target included in the set expressed by the search expression, an associative dictionary that stores the association value of a certain keyword to a certain keyword, and the degree to which a searcher associates a certain keyword with a certain keyword Has one or more rules that
A rule connection part that calculates an association value by connecting,
Based on the association value of the associative dictionary and the connection result in the rule connection part,
Find the association value from the input keyword to the additional keyword,
Based on the priority calculated by the priority calculation unit, the priority calculation unit that calculates the priority from the association value and the search target included in the set expressed by the search formula, based on the priority calculated by the priority calculation unit. Outputs the search target according to the priority given by the search target comparison unit with respect to the search target output by the search execution unit, and the search target comparison unit that has a priority addition unit that assigns a higher priority to the target And an output unit that operates.

<Structure of Claim 9> According to the structure of Claim 9, in the information search apparatus according to Claim 8, one rule is that the degree to which a searcher associates a certain keyword with a certain keyword is its association value. As described above, it is characterized in that it is provided with a rule connecting section for connecting the numbers equal to or less than a certain threshold value and calculating the association value from the directed graph obtained first.

<Explanation of Claims 8 and 9> In the inventions of Claims 8 and 9, association is made by connecting one or more rules in which the degree to which a searcher associates a keyword with a certain keyword is its association value. Calculate the valuation and use the associated valuation to calculate and attach a higher priority to the search target included in the set represented by the search formula, as the search target intended by the searcher. Is characterized by.

When a searcher inputs a search expression from the input section,
The search execution unit searches the database based on the search expression and outputs the searched search target. Further, the rule concatenation unit in the search target comparison unit calculates the association value by concatenating one or more rules that the searcher associates a certain keyword with a degree of the association value. Then, the priority calculation unit obtains the association value from the input keyword to the additional keyword based on the association value of the association dictionary and the connection result of the rule connection unit, and calculates the priority from this association value. The priority adding unit assigns, to the search targets included in the set expressed by the search formula, a priority that is higher as the search target intended by the searcher, based on the priority calculated by the priority calculation unit. wear. The output unit outputs, for example, a search target whose priority is higher than a certain threshold.

Therefore, the associative dictionary is compared with the case where the priority is calculated and attached only from the associative value from the certain input keyword to the certain additional keyword stored in the associative dictionary (inventions of claims 4 and 5). If the same number of associative values are stored, a more appropriate priority can be given, and the effect that the searcher does not output an unintended search target in the search formula is further enhanced. be able to.

<Structure of Claim 10> In the information search apparatus according to claim 8, one or more rules that the degree to which a searcher associates a certain keyword with a certain keyword is its association value is a number less than a certain threshold value. , And each association value calculated from one or more directed graphs,
It is characterized by having a rule connection unit for calculating an association value.

<Explanation of Claim 10> According to the invention of Claim 10, the rule connecting part of the invention of Claim 9 sets a rule that the degree to which a searcher associates a certain keyword with a certain keyword is its associated value. While associative value is calculated from the first obtained directed graph by connecting two or more, the rule is connected by a number less than a certain threshold, and the maximum value of each associative value calculated from one or more associative values or The feature is that an association value is calculated from an average value or the like. This allows
Similar to the invention of claim 9, a more appropriate association value can be obtained.

<Structure of Claim 11> In the information retrieving apparatus according to claim 8, one or more rules that the degree to which a searcher associates a certain keyword with a certain keyword is its association value are associated with a certain node. It is characterized in that it has a rule connection unit that calculates associative value from each associative value calculated from one or more directed graphs until all rules stored in the dictionary starting from the same node are connected. To do.

<Explanation of Claim 11> The invention of Claim 11 is the same as the rule connection part of the inventions of Claims 9 and 10 allows only one type of rule connection starting from a certain node. It is characterized by connecting all the rules starting from a node. Then, the maximum value, the average value, and the like of the associative values calculated from all the digraphs that can be connected are set as the associative value from the input keyword to the additional keyword. As a result, a more appropriate association value can be obtained as in the case of the invention of claim 9.

<Structure of Claim 12> An input unit for inputting a search expression, a database for storing a search target, and a search for the search target in the database are executed based on the search expression input from the input unit. , A search execution unit that outputs a search target included in a set expressed by a search expression, an associative dictionary that stores an association value from a keyword to a keyword, and a keyword set calculation unit that calculates a subset of a set And a rule concatenation unit that calculates an association value by connecting one or more rules that a searcher associates a keyword with a keyword with the association value. Associative value from a subset of the set of input keywords to a subset of the set of additional keywords, based on the subset obtained by the calculation unit and the connection result by the rule connection unit Based on the priority calculated by the priority calculation unit that calculates the priority from this association value and the search target included in the set expressed by the search formula, the searcher intends The search target comparison unit having a priority addition unit that gives a higher priority to the search target and the search target output from the search execution unit according to the priority assigned by the search target comparison unit. Is provided.

<Structure of Claim 13> An input unit for inputting a search expression, a database for storing the search target, and a search for the search target in the database are executed based on the search expression input from the input unit. , A search execution unit that outputs a search target included in a set expressed by a search expression, an associative dictionary that stores an association value from a keyword to a keyword, and a keyword set calculation unit that calculates a subset of a set And one or more rules that a searcher associates a keyword with a certain keyword is the associative value, and the rule used for the concatenation is that a keyword from a subset of one keyword set to another keyword In the rule concatenation part that calculates the association value by calculating it from the association value to the subset of the set of, and the association value and keyword set calculation part of the association dictionary Priority calculation that calculates the association value from the subset of the input keyword set to the subset of the set of additional keywords based on Based on the priority calculated by the priority calculation part, the search target included in the set represented by the section and the search formula is given a higher priority than the search target intended by the searcher. A search target comparison unit having an addition unit, and an output unit for outputting the search target according to the priority given by the search target comparison unit with respect to the search target output by the search execution unit. It is a thing.

<Explanation of Claims 12 and 13> Claim 12 and
In the thirteenth invention, instead of determining whether or not the rules can be connected from the input keyword to the additional keyword in the inventions of claims 8 to 11, the rule from the subset of the set of input keywords to the subset of the set of additional keywords is determined. It is characterized in that it is determined whether or not can be connected. For example, in the invention of claim 12, the association value from a subset S1 of a set of certain input keywords to a subset S2 of a set of certain additional keywords is min (V1, V2, ..., Vn) -con.
Calculate as st, etc. Here, V1, V2, ..., Vn are associative values of rules for connecting S1 to S2. Therefore, the associative dictionary is compared with the case where the priority is calculated and attached only from the associative value to the certain additional keyword set stored in the associative dictionary (claims 6 and 7). If the same number of associative values are stored, a more appropriate priority can be given, and the effect that the searcher can not output an unintended search target in the search formula is further enhanced. be able to.

<Structure of Claim 14> An input unit for inputting a search formula, a database for storing the search target, and a search for the search target in the database are executed based on the search formula input from the input unit. , A search execution unit that outputs a search target included in a set represented by a search expression, an associative dictionary that stores an association value from a keyword to a keyword, and a synonym that stores a list of synonymous keywords Based on the dictionary, the associative dictionary and the synonym dictionary, one or more sets of keywords S1 'obtained by replacing some of the input keywords with synonyms are obtained from the set of input keywords S1 and additional keywords are added from the set of additional keywords S2. Obtain more than one set S2 'of keywords with some of them replaced by synonyms, and add a set of additional keywords from the set S1 of input keywords
Associative value to S2 is set as an associative value from the set of keywords S1 'to the set of keywords S2', and based on the associative value in the associative value complementing unit, a set of additional keywords To the search target included in the set expressed by the search formula, and the priority that is higher for the search target that the searcher intends is set. A search target comparison unit having a priority addition unit to be attached, and an output unit for outputting the search target according to the priority assigned to the search target comparison unit with respect to the search target output from the search execution unit. It is a feature.

<Explanation of Claim 14> According to the invention of Claim 14, a set of keywords S1 'in which some of the input keywords are replaced with synonyms from the set of input keywords S1'.
From the set S2 of additional keywords, one or more sets S2 'of keywords obtained by replacing some of the additional keywords with synonyms are obtained, and the association value from S1 to S2 is S1' The feature is that it is configured to be obtained as an association value to. Therefore, if the associative dictionary stores the same number of associative values as compared with the case (Claims 1 and 2) which does not do so, a more appropriate priority can be given and the searcher can use the search expression. It is possible to further enhance the effect of not outputting an unintended search target.

<Structure of Claim 15> An input unit for inputting a search formula, a database for storing a search target, and a search for the database are executed based on the search formula input from the input unit. A search execution unit that outputs a search target included in a set represented by a search expression, an associative dictionary that stores an association value from a keyword to a keyword, and a synonym dictionary that stores a degree of synonym of the keyword, Set of input keywords based on associative dictionary and synonym dictionary
A set of keywords S1 'in which some of the input keywords are replaced with synonyms, and a set of keywords S2' in which some of the additional keywords are replaced with synonyms are added to each of the keywords before and after replacement. Get as high a degree of synonym as possible, set of input keywords S1
From the associative value from the keyword set S1 'to the keyword set S2', the associative value from the keyword set S1 'to the additional keyword set S2, and the associative value synthesis unit that is obtained from the similarity of the keywords before and after the replacement,
Based on the associative value in the associative value synthesizer, find the associative value from the set of input keywords to the set of additional keywords, calculate the priority from this associative value, and search included in the set expressed by the search formula. The search target comparison unit has a priority addition unit that gives a higher priority to the target than the search target intended by the searcher, and the search target comparison is performed on the search target output from the search execution unit. An output unit that outputs a search target according to the priority given by the unit.

<Explanation of Claim 15> According to the invention of Claim 15, a set of keywords S1 'in which some of the input keywords are replaced with synonyms from the set of input keywords S1'.
And a set S2 ′ of keywords obtained by replacing some of the additional keywords with synonyms from the set S2 ′ of additional keywords.
Is the synonym of each keyword before and after replacement (general synonym)
Is as large as possible, and the association value from S1 to S2 is S
The feature is that it is obtained from the association value from 1'to S2 'and the synonym of the keywords before and after the replacement. Therefore,
If the associative dictionary stores the same number of associative values as compared with the case (Claims 1 and 2) which does not do so, a more appropriate priority can be given and the searcher intends in the search formula. It is possible to further enhance the effect of not outputting the search target that has not been output.

<Structure of Claim 16> An input unit for inputting a search expression, a database for storing a search target, and a search for the search target in the database are executed based on the search expression input from the input unit. , A search execution unit that outputs a search target included in a set expressed by a search expression, an associative dictionary that stores an association value from a keyword to a keyword, and a synonym that stores a list of synonymous keywords A dictionary and a synonym dictionary that stores the degree of similarity of keywords,
Based on the associative dictionary and the synonym dictionary, a first set of replacement keywords obtained by replacing some of the keywords with synonyms from the first set of keywords, and some of the keywords from the second set of synonyms To obtain a second set of replacement keywords such that the synonyms of the keywords before and after replacement are as large as possible, and the association value from the first set of keywords to the second set of keywords is An associative value synthesizing unit that obtains the association value from the set of replacement keywords to the second set of replacement keywords and the similarity of the keywords before and after the replacement, and some of the keywords from the first set of replacement keywords as synonyms. To a third set of replacement keywords and a fourth set of second replacement keywords in which some of the keywords are replaced with synonyms. Obtained one or more of the set of conversion keyword,
An associative value complementing unit that sets the associative value from the first keyword set to the second keyword set as the associative value from the third replacement keyword set to the fourth replacement keyword set;
Based on the associative value in the associative value complementing unit, obtain the associative value from the set of input keywords to the set of additional keywords, calculate the priority from this associative value, and search included in the set expressed by the search formula. The search target comparison unit has a priority addition unit that gives a higher priority to the target than the search target intended by the searcher, and the search target comparison is performed with respect to the search target output from the search execution unit. An output unit that outputs a search target according to the priority given by the unit.

<Explanation of Claim 16> The invention of claim 16 is a set of input keywords (first set of keywords).
Of the input keywords from S1, some of the input keywords are replaced with synonyms (first set of replacement keywords) S1 ′, and of the set of additional keywords (second set of keywords) S2 , A set of keywords (second set of replacement keywords) S2 ′ in which some of them are replaced with synonyms are obtained so that the degree of synonym (comprehensive degree of synonym) of each keyword before and after the substitution is as large as possible.
Of the 1'keywords, a set of keywords (third set of keywords) S1 "in which some are replaced with synonyms,
A set of keywords obtained by replacing some of the keywords of S2 'with synonyms (a set of fourth keywords) S2 ", and a set of additional keywords from the set S1 of input keywords
The feature is that the associative value to S2 is calculated from the associative value from the third replacement keyword set S1 ″ to the fourth replacement keyword set S2 ″. That is, the present invention has the configurations of claims 14 and 15.

Therefore, as compared with claim 14, if the associative dictionary stores the same number of associative values, it is possible to give a more appropriate priority, which is not intended by the searcher in the search formula. The effect of not outputting the search target can be further enhanced.

<Structure of Claim 17> When the input part for inputting the search expression, the database for storing the search target, and the search expression input from the input part are ambiguous search expressions Based on the search formula change part that converts to a search formula whose intention is clear, and the search formula output by the search formula change part, the search target of the database is executed and included in the set expressed by the search formula. Included in the set expressed by the search formula, using the search execution unit that outputs the search target that is output, and the association value from the set of input keywords that are included in the search formula to the set of additional keywords that are added to the search target. The search target comparison unit that gives a higher priority to the search target that the searcher intends, and the search target comparison unit assigns to the search target output from the search execution unit. Search according to priority It is characterized in that an output unit for outputting an elephant.

<Explanation of Claim 17> The invention of Claim 17 is characterized in that a search expression changing unit is added to the configuration of the invention of Claim 1. The search expression changing unit, when the search expression input by the searcher is a search expression whose intention is ambiguous,
Convert to a search expression that has a clear intention. Therefore, as compared with the case (Claims 1 to 16), it is possible to prevent the searcher from outputting an unintended search target in the search expression, and the search that the searcher intended in the search expression. The effect that the target can be output can be obtained.

<Structure of Claim 18> When the input part for inputting the search expression, the database for storing the search target, and the search expression input from the input part are ambiguous search expressions Based on the search formula change part that converts to a search formula whose intention is clear, and the search formula output by the search formula change part, the search target of the database is executed and included in the set expressed by the search formula. Included in the set expressed by the search formula using the association value from the input keyword to the additional keyword included in the search formula whose intention is output by the search execution unit that outputs the search target The search target comparison unit that gives a higher priority to the search target as the search target is intended by the searcher, and the priority given by the search target comparison unit to the search target output from the search execution unit. Search target according to degree It is characterized in that an output unit for outputting.

<Explanation of Claim 18> In the invention of Claim 18, in addition to the invention of Claim 17, the search expression changing unit outputs one or more search expressions whose intention is clear, and these search expressions are expressed. It is characterized in that the search targets included in the set are given priority. Therefore, the search target intended by the searcher can be output without omission, as compared with the invention of claim 17.

<Structure of Claim 19> When the input part for inputting the search expression, the database for storing the search target, and the search expression input from the input part are ambiguous search expressions Based on the search formula change part that converts to a search formula whose intention is clear, and the search formula output by the search formula change part, the search target of the database is executed and included in the set expressed by the search formula. Included in the set expressed by the search formula using the association value from the input keyword to the additional keyword included in the search formula whose intention is output by the search execution unit that outputs the search target A search target comparison unit that gives a higher priority to a search target that the searcher intends for the search target, a priority for the converted search formula output by the search formula change unit, and a search target comparison unit Search target output by And overall priority calculation unit to give an overall priority reflects both priority against, is characterized in that an output unit that outputs the search target according overall priorities.

<Explanation of Claim 19> According to the invention of Claim 19, with respect to the configuration of Claims 17 and 18, for a search target included in a set expressed by a search expression having a clear intention,
Comprehensive priority calculation that gives priority (overall priority) that reflects both the priority of the converted search expression output by the search expression changing unit and the priority of the search target output by the search target comparison unit It is characterized by the addition of parts. Therefore, as compared with the case where such a process is not performed (Claims 1 to 18), it is possible to prevent the searcher from outputting an unintended search target in the search formula, and the searcher can search. The effect that the search target intended by the expression can be output can be made larger.

<Structure of Claim 20> An input unit for inputting a search expression, a database for storing a search target, and a search for the search target in the database are executed based on the search expression input from the input unit. A search execution unit that outputs the search target included in the set expressed by the search formula, and a search result division unit that divides the search target included in the set expressed by the search formula into one or more sets , For each set of search targets divided by the search result division unit, the search formula is expressed using the association value from the set of input keywords included in the search formula to the set of additional keywords added to the search target. The search target that the searcher intends, the search target comparison unit that gives a higher priority to the search target included in the set, and the search target is output according to the priority assigned by the search target comparison unit. Output section The is characterized in that it comprises.

<Explanation of Claim 20> According to the invention of Claim 20, in the structure of the invention of Claims 1 and 2, the search target included in the set represented by the search expression is divided into one or more sets. It is characterized by the addition of a search result division unit that does. Therefore, the search target comparison unit executes the search target comparison processing for each of the divided search target sets, and assigns a priority. Then, the output unit outputs the search target according to this priority. Therefore, as compared with the case where such division processing is not performed (claims 1 to 19), there is an effect that the time until the device starts responding to the searcher can be shortened.

<Structure of Claim 21> An input unit for inputting a search expression, a database for storing the search target, and a search for the search target in the database are executed based on the search expression input from the input unit. , A search execution unit that outputs the search target included in the set expressed by the search expression, and the search result division that divides the search target included in the set expressed by the search expression into one or more search target sets Part, a process distribution unit that distributes the search target comparison process for each search target set divided by the search result dividing unit, and the search target comparison process that is distributed by the process distribution unit, respectively executed in parallel, and input keyword One or more searches that give a higher numerical priority to the search target that the searcher intends for the search target that is included in the set expressed by the search expression, using the association value from Target ratio And parts, and is characterized in that an output unit that outputs the search target in accordance with the priority of the search target comparison unit.

<Explanation of Claim 21> In the invention of Claim 21, as in the invention of Claim 20, the search result division unit divides into one or more retrieval object sets, and further, the processing distribution unit performs retrieval. The feature is that search target comparison processing for each set of one or more targets is assigned to one or more search target comparison units. The search target comparison process is executed in parallel by one or more search target comparison units.
As compared with the invention of (1), it is possible to obtain the effect that the time until the device finishes responding to the searcher can be further shortened.

[0053]

BEST MODE FOR CARRYING OUT THE INVENTION The information retrieval device of the present invention is constructed according to the following principle. That is, the more the search target intended by the searcher, the easier it is for the additional keyword of the search target to be associated with the input keyword that is the keyword included in the search formula. The numerical value indicating the strength of this tendency is called the association value. Therefore, by using the associative value, it is possible to calculate and give priority to a search target included in the set expressed by the search formula, which has a higher numerical value for the search target intended by the searcher. By outputting the search target based on the degree, the search target intended by the searcher can be output.

In the phenomenon of human thinking called association, for example, if a set A of keywords is associated with a set B of keywords and a set B of keywords is associated with a set C of keywords, the set A of keywords is From the above, there is a phenomenon that the set C of keywords is associated. Focusing on these phenomena, it is possible to calculate other association values from one association value. In addition, since linguistic knowledge such as synonyms and synonyms is deeply related to phenomena in human thinking such as association, it can be used for calculation of association value. For example, keyword A and keyword A'are synonyms and keyword A
If the association value from V to B is V, then keyword A '
The associative value from B to B is also V.

Based on such a principle, the present invention uses the association value from the input keyword to the additional keyword and the linguistic knowledge such as synonyms and synonyms to search objects included in the set represented by the search expression. On the other hand, the higher the search target that the searcher intends, the higher the priority is calculated so that the searcher does not output the search target that was not intended by the search formula, and the intention is clear. After changing the search formula so that it becomes, the searcher outputs the search target intended by the search formula by calculating the priority for the search target as described above. By dividing and executing the priority calculation, the time required for the device to respond is shortened.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. << Specific Example 1 >><Configuration> FIG. 1 is a configuration diagram of a specific example 1 in the information retrieval apparatus of the present invention. In the figure, 101 is an input unit for a searcher to input a search expression, and 102 is a database 103.
A search execution unit that executes a search to the search target of and outputs the search target included in the set represented by the search expression, 103.
Is a database that stores search targets such as documents, 1
Reference numeral 04 represents the association value from the set of input keywords to the set of additional keywords as a priority that is higher in numerical value than the search target intended by the searcher with respect to the search target included in the set expressed by the search formula. Search target comparison unit to be attached, 105
For the search target output by the search execution unit 102,
The output unit outputs the search target according to the priority given by the search target comparison unit 104.

In the following description, the search expression input to the input unit 101 is a search expression expressed in Boolean algebra.
SQ used for input to predicate logic and relational database
It may be a search expression expressed by another expression method such as L (Structured Query Language).

FIG. 2 is a diagram showing the configuration of the search object comparison unit 104. In the figure, 201 is an associative dictionary 202 that indicates an association value from a set of input keywords to a set of additional keywords.
The priority addition unit 202 obtains the association value as a priority with respect to the search target included in the set expressed by the search expression, and 202 adds from the set of certain keywords to the set of certain keywords. It is an associative dictionary that stores the associative value of.

FIG. 3 is a diagram showing a specific example of the associative dictionary 202. For example, an association value from a set of keywords (association source keyword set) {"computer", "word processor"} to a set of keywords (association destination keyword set) {"computer", "outline font"} is 400
It is. Further, the association value from a certain association source keyword set to the same association destination keyword set is not stored in order to save the storage capacity of the association dictionary, and the association value is 100.

<Operation> FIG. 4 is a processing flowchart of the first specific example. First, in step 401, the search execution unit 102 executes a search on the database 103 to obtain a search target included in the set represented by the search expression. In step 402, the search target comparison unit 104 assigns the association value from the set of input keywords to the set of additional keywords to the search targets included in the set represented by the search expression as a priority (search). Target comparison processing).
In step 403, the output unit 105 outputs the search target according to the priority. For example, a search target whose priority is higher than a certain threshold is output.

FIG. 5 is a flowchart of the process in step 402 (search target comparison process). First, in step 501, an iterative process that continues until step 504 in which search targets are used one by one is started. Step 5
In 02, the association value from the set of input keywords to the set of additional keywords is obtained (all association value search processing). In step 503, the priority is obtained from the association value by setting the association value as the priority.

FIG. 6 is a flowchart of the process in step 502 (all associative value search process). First, in step 601, the associative dictionary 202 is accessed,
If the associative dictionary 202 stores the associative value from the set of input keywords to the set of additional keywords, the associative value is obtained and the process ends. If not, the process proceeds to step 602. In step 602, the association value from the set of input keywords to the set of additional keywords is set to zero.

An example of the operation of the above-described specific example 1 will be briefly described below. First, it is assumed that the searcher inputs the search expression "computer or word processor" with the intention of "general information about computers and word processors." In addition, the contents of the database 103 are DTP (DeskTop Publishing),
Reference 1 in which the additional keywords are “computer” and “outline font”, and the content is ultra-high-speed operation and the additional keywords are “computer” and “multiprocessor”
Suppose there is a reference 2 that is. Document 1 is the search target intended by the searcher, but document 2 is not. Step 4
In 01, documents 1 and 2 are obtained as search targets. In step 402, the priorities of Document 1 and Document 2 are obtained.

FIG. 7 shows the association value obtained in step 502. As shown, the priority of Document 1 is 400,
The priority of Document 2 is 0. In step 403,
If the threshold value of the priority of the output search target is 300, only the document 1 that is the search target intended by the searcher is output.

<Effect> As described above, according to the first specific example, the associative value from the set of input keywords to the set of additional keywords is set for the search target included in the set represented by the search expression. Since the search target intended by the searcher is given a higher priority as a search target, it is possible to prevent the searcher from outputting a search target that was not intended by the search formula.

<Specific Example 2><Structure> Since the structure of the specific example 2 is the same except the internal structure of the search target comparison unit 104 in the specific example 1, it will be described with reference to FIG.

In the second specific example, the search target comparison unit 104
Selects a set of input keywords and a set of additional keywords that do not affect the result of the search by the search execution unit 102, and obtains the association value from the set of input keywords to the set of the additional keywords as a search expression. With respect to the search targets included in the represented set, the search targets intended by the searcher are configured to have a higher numerical priority.

FIG. 8 is a diagram showing an internal configuration of the search target comparison unit 104 in the second specific example. In the figure, reference numeral 801 indicates, using the associative dictionary 803, the association value from the set of input keywords selected by the keyword selection unit 802 to the set of additional keywords, and the search expression expresses this association value. A priority addition unit 802 that gives a higher priority to a search target intended by the searcher than a search target included in the set, and an input 802 that does not affect the search result of the search execution unit 102. A keyword selection unit for selecting a set of keywords and a set of additional keywords, 80
Reference numeral 3 is an associative dictionary that stores association values from a certain keyword set to a certain keyword set.

Here, if a search expression includes a search target in a set and the search expression includes the same search target even if the input keyword is removed from the search expression, the input keyword is the search execution unit 102. It is an input keyword that does not affect the results of the search in. Further, even if a certain additional keyword is excluded from the search target, if the search formula includes the same search target, the additional keyword is an additional keyword that does not affect the result of the search by the search execution unit 102.

By the way, the search objects included in the set represented by the search formula have a commonality that they are included in the set of the search formula. Therefore, in order to make the difference in priority appear large, the set of input keywords and the additional keywords that do not affect the commonality, that is, do not affect the result of the search by the search execution unit 102, are added. It is better to use a set. Search target comparison unit 104 of this specific example
Is for performing such search target comparison processing.

<Operation> The operation of the second specific example is the same except for step 402 (search target comparison processing) in the first specific example. The search target comparison process of the specific example 2 is different from the search target comparison process of the specific example 1 in Step 5 of the specific example 1.
It is the same except that the keyword selection process is performed immediately before 02 (all associative value search process).

FIG. 9 is a flowchart of the search target comparison processing of the second specific example. In step 902, a set of input keywords and a set of additional keywords that do not affect the result of the search by the search execution unit 102 are selected (keyword selection process). Operation other than this step is specific example 1
Therefore, the description is omitted.

An example of the operation of the second specific example will be briefly described below. In the iterative process relating to Document 1, in the keyword selection process, {“word processor”} and {“outline font” are set as a set of input keywords and a set of additional keywords that do not affect the result of the search by the search execution unit 102, respectively. } Is selected. In the iterative process related to the document 2, {“word processor”} and {“multiprocessor”} are selected, respectively.

<Effect> As described above, according to the second specific example, a set of input keywords and a set of additional keywords that do not affect the result of the search by the search execution unit 102 are selected, and the set of the input keywords is selected. Is more appropriate than the case where the association value from the set of input keywords to the set of additional keywords is the priority (specific example 1). It is possible to obtain a greater effect that it is possible to prevent the searcher from outputting an unintended search target in the search expression.

<Specific Example 3><Structure> Since the structure of the specific example 3 is the same except the internal structure of the search target comparison unit 104 in the specific example 1, the description will be given again with reference to FIG. The search target comparison unit 104 of the third specific example uses the association value from each input keyword to each additional keyword, and the searcher intends for the search target included in the set represented by the search formula. The feature is that the higher the numerical value is, the higher the priority is calculated for the search target.

FIG. 10 is a diagram showing the internal structure of the search object comparison unit 104 in the third specific example. In the figure, 1001
Is based on the priority calculated by the priority calculation unit 1002 for the search target included in the set represented by the search formula,
A priority addition unit 1002 that gives a higher priority to a search target intended by a searcher is a priority calculation unit 1002 that calculates a priority from an association value of each input keyword to each additional keyword, and 1003 is a priority calculation unit 1003. , Is an associative dictionary that stores association values from a certain keyword to a certain keyword.

FIG. 11 is a diagram showing a specific example of the associative dictionary 1003. For example, keyword (association source keyword)
The association value from the "computer" to the keyword (association destination keyword) "outline font" is 60.

<Operation> The operation of the specific example 3 is the same as step 502 (all associative value search processing) and step 5 in the specific example 1.
Other than 03 is the same.

FIG. 12 is a flowchart of the all-associative-value search process of the third specific example. First, in step 1201, step 120 using input keywords K1 one by one
The repeated processing up to 5 is started. Next, step 1
In 202, the iterative process that continues to step 1204 using the additional keywords K2 one by one is started. In step 1203, an association value from the input keyword K1 to the additional keyword K2 is obtained (each association value search process).

FIG. 13 is a flow chart of the processing in step 1203 (each association value search processing). First,
In step 1301, the associative dictionary 1003 is accessed, and if the associative dictionary 1003 stores the associative value from the input keyword K1 to the additional keyword K2, the associative value is obtained and the processing ends. If not, step 1302. Go to. In step 1302, the association value from the input keyword K1 to the additional keyword K2 is set to zero. In step 503, the priority is calculated from the association value. In this calculation, for example, the sign of the association value between the input keyword to which the negation operator is combined and a certain additional keyword is reversed, and then the sum of the obtained association values is set as the priority. Calculate according to the method.

An example of the operation of Concrete Example 3 will be briefly described below.
In the iterative process from step 501 to step 504 shown in FIG. 5, first, the process is performed on the document 1 which is the first search target.

FIG. 14 shows a step 50 regarding the document 1.
It is a figure which shows the association value obtained by 2. In step 503, the priority of Document 1 is 100 + 60 + 90 + 95 = 3.
It becomes 45. Similarly, the priority of the document 2 which is the second search target is 100 + 70 + 90 + 0 = 260.

Further, in the third specific example, in order to speed up the process (each association value search process) of obtaining an association value from a certain input keyword to a certain additional keyword, the retrieval object comparison unit 104 is made to cache the association value. By providing the memory, it is possible to obtain the associative value from the associative value cache memory instead of from the associative dictionary 1003.

<Effect> As described above, according to the third specific example, the associative value from each input keyword to each additional keyword is used to search for objects included in the set expressed by the search expression. So, the more search target the searcher intends,
In order to calculate the priority which is a high numerical value, as in the first specific example, as compared with the case where the association value from the set of input keywords stored in the associative dictionary 202 to the set of additional keywords is set as the priority, If the associative dictionary 1003 stores the same number of associative values, a more appropriate priority can be given, and the searcher can be prevented from outputting an unintended search target in the search formula. You can get more.

<Specific Example 4><Structure> The structure of Specific Example 4 is the same except for the internal structure of the search target comparison unit 104 in Specific Example 1. The search target comparison unit 104 of the fourth specific example uses the associative value from the subset of the set of input keywords to the subset of the set of additional keywords, with respect to the search target included in the set represented by the search expression. The search target intended by the searcher is configured to calculate and attach a priority having a higher numerical value.

FIG. 15 is an internal block diagram of the search object comparison unit 104 in the fourth specific example. In the figure, reference numeral 1501 indicates a priority that is higher for a search target intended by the searcher, based on the priority calculated by the priority calculation unit 1502, for the search target included in the set expressed by the search formula. Priority addition unit 1502, a keyword set calculation unit 1520
The priority calculation unit 1503 that obtains the association value from the subset of the set of input keywords calculated in 03 to the subset of the set of additional keywords based on the association dictionary 1504, and calculates the priority from the association value, A keyword set calculation unit 1504 for calculating a subset of a certain set is an associative dictionary that stores an association value from a certain keyword set to a certain keyword set.

<Operation> The operation of the fourth specific example is the same except for step 402 (search target comparison process) in the first specific example.

FIG. 16 is a flowchart of the search target comparison process of the fourth specific example. First, in step 1601, a power set of a set of input keywords is calculated. In step 1602, the iterative process that continues to step 1605 using search targets one by one is started. In step 1603, an association value is obtained from the association dictionary 1504 for each combination of a subset of the set of input keywords and a subset of the set of additional keywords (all association value search processing).
In step 1504, the priority is calculated from the association value by the same method as the operation of the third specific example.

FIG. 17 is a flowchart of the process in step 1603 (all-association-value search process). First,
In step 1701, a power set of a set of additional keywords is calculated. In step 1702, the iterative process that continues until step 1706 in which one of the subsets S1 of the set of input keywords other than the empty set is used is started. In step 1703, the iterative process that continues until step 1705 in which one of the subsets S2 of the set of additional keywords other than the empty set is used is started. In step 1704, the associative value from S1 to S2 is obtained (each associative value searching process).

FIG. 18 is a flowchart of the process (each associated value search process) in step 1704. First,
In step 1801, the associative dictionary 1504 is accessed, and if the associative dictionary 1504 stores the associative value from S1 to S2, the associative value is obtained and the processing ends. If not, the process proceeds to step 1802. In step 1802,
Associativity from S1 to S2 is zero.

An example of the operation of Concrete Example 4 will be briefly described below. FIG. 19 shows an example of the associative value stored in the associative dictionary 1504. In step 1601 above, a set of input keywords {“computer”, “word processor”}
Power set {{"computer", "word processor"},
Get {"computer"}, {"word processor"}, {}}. In the iterative process from step 1602 to step 1605, first, the process regarding the document 1 which is the first search target is performed.

FIG. 20 is a diagram showing the association value obtained in step 1603. In step 1604, the priority is 180 + 100 + 90 = 370. Similarly, the priority of the document 2 which is the second search target is 100 + 70.
It becomes + 90 = 260.

In the fourth specific example, if the keyword set calculation unit 1503 does not calculate a subset of a certain set but calculates a set having all or some of the elements of a certain set, It is also possible to calculate the priority by using the association value from the set of keywords including all or some of the elements of the set of input keywords to the set including all or some of the elements of the set of additional keywords. . Accordingly, for example, even when the association dictionary 1504 does not store the associative value of a subset of a certain set, the associative value can be set by appropriately setting a keyword set including all or some of the elements of the set. Can be calculated.

<Effect> The expressiveness of the association value is greater when it is expressed by the association value from a certain keyword set to a certain keyword set than by expressing it by only the association value from a certain keyword to a certain keyword. Therefore, according to the specific example 4, since the priority is calculated using the association value from the subset of the set of input keywords to the subset of the set of additional keywords, the association value from a certain input keyword to a certain additional keyword is calculated. If the associative dictionary stores the same number of associative values as compared to the case where the priority is calculated from only (specific example 3), a more appropriate priority can be given and the searcher can use the search expression. It is possible to obtain the effect that an unintended search target can be prevented from being output.

<Specific Example 5><Structure> The structure of the specific example 5 is the same except for the internal structure of the search target comparison unit 104 in the specific example 1. In the fifth specific example, the search target comparison unit 104 calculates the association value by connecting one or more rules that the degree to which a searcher associates a keyword from a certain keyword is the association value, and the association value is calculated. Using, the search target included in the set represented by the search formula is calculated and given a higher priority as the search target intended by the searcher.

FIG. 21 is an internal block diagram of the search object comparison unit 104 in the fifth specific example. In the figure, reference numeral 2101 denotes a priority that is higher for a search target intended by the searcher, based on the priority calculated by the priority calculation unit 2102, for the search target included in the set expressed by the search formula. The priority addition unit 2102 that attaches a priority calculation unit 2103 that calculates the priority from the association value from the input keyword to the additional keyword based on the association result of the rule connection unit 2103 and the association value of the association dictionary 2104. Is a rule concatenation unit 2104 that calculates an association value by connecting one or more rules that a searcher associates one keyword with a degree that the association value is the association value. It is an association dictionary that stores association values.

Here, if one or more rules that the degree to which a searcher associates a certain keyword with a certain keyword is the association value are connected, a node is a keyword,
This is a relation in which arcs are associated, and a directed graph corresponding to one associative value is formed in each arc.

<Operation> The operation of the fifth specific example is the same except for step 1203 (each associated value search process) in the third specific example.

FIG. 22 is a flowchart of each associative value search process of the fifth specific example. First, in step 2201, the iterative process that continues until step 2203, which is performed while increasing the number of keyword chain arcs from 1 to a certain threshold, is started. In step 2202, it is determined whether the rule can be connected from the input keyword K1 to the additional keyword K2 based on the number of arcs, and if possible, step 220
If not possible, go to step 2203. Here, the rule can be connected from K1 to K2 means that a directed graph having K1 as a starting point and K2 as an ending point can be formed by using the rules stored in the associative dictionary 2104. Further, this connection is performed by a well-known algorithm such as vertical search or horizontal search. In step 2204, the association value from K1 to K2 is set to zero. Step 220
In 5, the association value from K1 to K2 is calculated from the formed directed graph. In this calculation, for example, the associative value from K1 to K2 is a value obtained by subtracting a certain value from the minimum value of the associative value corresponding to the arc passing from K1 to K2 in the directed graph. Calculate by the method of.

An example of the operation of Concrete Example 5 will be briefly described below. FIG. 23 is an example of the association value stored in the association dictionary 2104. From the input keyword "computer",
Assume that the association value for the additional keyword "outline font" is obtained. In step 2201, the number of arcs in the keyword chain is set to 1. In step 2202, when the number of arcs is 1, the association dictionary 2104 does not store the rule that the association source keyword is “computer” and the association destination keyword is “outline font”. ”
Is not possible, and the processing proceeds to step 2203. If the threshold value is 5, the process proceeds to step 2201 and the number of arcs is 2
And In step 2202, if the number of arcs is 2,
A directed graph having “computer” as a starting point and “outline font” as an ending point can be formed and can be connected, and the process proceeds to step 2205.

FIG. 24 is a diagram showing the directed graph. In step 2205, if the value subtracted from the minimum association value is 20, the association value from "computer" to "outline font" is min (90,95)-
20 = 70.

Further, in the fifth specific example, in order to obtain a more appropriate associative value, even if the rule can be connected from the input keyword K1 to the additional keyword K2, the process of connecting the rules is performed by the number of arcs. It is also possible to continue until a certain threshold is reached, and use the maximum value or average value of the associative values calculated from one or more directed graphs as the associative value from the input keyword K1 to the additional keyword K2.

Further, in the fifth embodiment, in order to obtain a more appropriate associative value, even if the rule can be connected from the input keyword K1 to the additional keyword K2, the associative dictionary 2104 is stored at a certain node. It continues until all the rules that have the same node as the starting point are connected, and the maximum value or the average value of each associative value calculated from one or more directed graphs is calculated from the input keyword K1 to the additional keyword K2. It is also practicable. That is, when there are a plurality of directed graphs having the same node as the starting point, the associative value may be calculated from each directed graph.

<Effect> As described above, according to the fifth specific example, one or more rules that the degree to which a searcher associates a certain keyword with a certain keyword is the association value,
Since the association value is calculated by linking, the priority is calculated and attached only from the association value from the certain input keyword stored in the associative dictionary to the certain additional keyword (specific example 3).
Compared with, if the associative dictionary stores the same number of associative values, it is possible to give more appropriate priority and prevent the searcher from outputting unintended search targets in the search formula. The effect can be further enhanced.

<Specific Example 6><Structure> The structure of the specific example 6 is the same except for the internal structure of the search target comparison unit 104 in the specific example 1. The search target comparison unit 104 of the sixth example calculates the association value by connecting one or more rules that the degree of association of a keyword set by a searcher with a keyword set is the association value, and the association value is calculated. By using the valence, the search object included in the set expressed by the search formula is calculated and given a higher priority as the search object intended by the searcher.

The internal structure of the search target comparison unit 104 in the sixth example is such that an associative dictionary that stores an association value from a keyword to a keyword, a keyword set calculation unit that calculates a subset of a set, and a searcher It was calculated by the rule connection part that calculates the association value by connecting one or more rules that the degree of associating a keyword from a keyword is the association value, and the association value and keyword set calculation part of the association dictionary. A priority calculation unit that calculates an association value from a subset of the set of input keywords to a subset of the set of additional keywords based on the subset and the connection result in the rule connection unit, and calculates a priority from this association value. , The higher the numerical value of the search target intended by the searcher, based on the priority calculated by the priority calculation unit, with respect to the search target included in the set expressed by the search formula It is made of a priority adding section to give a certain priority.

<Operation> The operation of the sixth specific example is the same except for step 1704 (each associated value search process) in the fourth specific example. Each associative value search process of the specific example 6 is such that, in step 2202 of each associative value search process of the specific example 5, instead of determining whether the rule can be connected from the input keyword K1 to the additional keyword K2, a set of input keywords is set. It is determined whether rules can be connected from the subset S1 of to the subset S2 of the set of additional keywords.

FIG. 25 is a diagram showing a directed graph obtained by connecting rules from a subset S1 of a set of certain input keywords to a subset S2 of a set of additional keywords. In this case, the association value from S1 to S2 is min (V1, V2, V
3) -Const etc.

Further, in the above specific example 6, in order to obtain a more appropriate association value, a subset of the set of input keywords
It is also possible to further obtain the rule used in the connection of the rules from S1 to the subset S2 of the set of additional keywords by the all-associative-value search process of the fourth specific example.

FIG. 26 is an explanatory diagram of a modified example of the sixth specific example. This figure shows that the keyword set A to the keyword set B, the association value from the subset A1 of the set A to the subset B1 of the set B, and the association value from the subset A2 of the set A to the subset B2 of the set B. It is a figure which shows the directed graph obtained by connecting using. In this case, the association value from the set A to the set B is calculated as V1 + V2.

<Effect> As described above, according to the sixth example, one or more rules are connected by associating one or more rules that the degree to which a searcher associates a certain keyword set with a certain keyword set is the association value. The value is calculated, and the priority is calculated from the association value from the subset of the set of input keywords to the subset of the set of additional keywords. If the associative dictionary stores the same number of association values as compared to the case where the priority is calculated and attached only to the set of keywords (concrete example 4), a more appropriate priority is given. Therefore, it is possible to further enhance the effect that the searcher can prevent an unintended search target from being output in the search formula.

<Specific Example 7><Structure> The structure of the specific example 7 is the same except for the internal structure of the search target comparison unit 104 in the specific example 1. Using the synonym dictionary, the search target comparison unit 104 of Specific Example 7 obtains at least one keyword set S1 ′ obtained by replacing some of the input keywords with synonyms using the synonym dictionary, and adds the additional keywords. Of the additional keywords from the set S2 of
, And obtain the association value from the set S1 of input keywords to the set S2 of additional keywords as the association value from the set S1 'of keywords to the set S2' of keywords. With respect to the search target included in, the search target intended by the searcher is given a higher priority.

FIG. 27 is an internal block diagram of the search object comparison unit in the seventh specific example. In the figure, 2701 is an associative dictionary 2.
Based on the association value in 703 and the association value in the association value complementing unit 2702, the association value from the set of input keywords to the set of additional keywords is calculated, and the priority is calculated from this association value.
A priority addition unit 2702 that gives a higher priority to a search target intended by a searcher than a search target included in the set represented by the search formula is a synonym dictionary 27.
Based on 04, one or more sets of keywords S1 'obtained by replacing some of the input keywords with synonyms are obtained from the set of input keywords S1 and some of the additional keywords are replaced with synonyms from the set of additional keywords S2 Associative value complementer that obtains one or more of the set S2 'of keywords and sets the association value from the set S1 of input keywords to the set S2 of additional keywords as the association value from the set S1' of keywords to the set S2 'of keywords , 2703 is an associative dictionary that stores association values from a certain keyword set to a certain keyword set, 2
A synonym dictionary 704 stores a list of keywords that are synonyms.

FIG. 28 is a diagram showing a specific example of the synonym dictionary 2704. For example, "computer", "computer", and "computer" are synonyms. The absence of synonyms for a keyword is expressed by not storing a list of synonyms including the keyword. For example, in FIG. 28, since a list including "outline font" is not stored, a synonym of "outline font" does not exist.

<Operation> The operation of the specific example 7 is different only in the all associative value search process of step 502 in the operation of the specific example 1.

FIG. 29 is a flowchart of the all-associative value search process of the seventh specific example. First, in step 2901, the synonym dictionary 2704 is accessed to obtain one or more keyword sets S1 ′ in which some of the input keywords are replaced with synonyms from the input keyword set S1 and the additional keyword set S2 is obtained. To obtain one or more keyword sets S2 ′ in which some of the additional keywords are replaced with synonyms. Where S1 is one of S1 'and S2 is S2'
Is one of the. In step 2902, S1 '
The iterative process that continues up to step 2906 using one by one is started. S1 'used first in this iterative process
Is S1. In step 2903, the iterative process that continues to step 2905 using S2 ′ one by one is started. S2 ′ used first in this iterative process is S2. In step 2904, the associative dictionary 2703 is accessed, and if the associative dictionary 2703 stores the associative value from S1 ′ to S2 ′, the process proceeds to step 2908, and if not, the process proceeds to step 2905. Step 29
At 07, the association value from S1 to S2 is set to zero. In step 2908, the association value from S1 'to S2' is set to the association value from S1 to S2.

An example of the operation of Concrete Example 7 will be briefly described below. FIG. 30 is an example of the association value stored in the association dictionary 2703. Assume that the association value from S1 = {“computer”, “word processor”} to S2 = {“computer”, “outline font”} is obtained. In step 2901, as S1 ', S1 and {"computer", "word processor"}, {"computer", "word processor"}
Then, as S2 ′, S2 and {“calculator”, “outline font”}, {“computer”, “outline font”} are obtained. Steps 2902 to 290
When S1 ′ = {“calculator”, “word processor”} and S2 ′ = {“calculator”, “outline font”} in the iterative process continuing up to 6, the associative dictionary 2703 changes from S1 ′ to S2 ′. Remember the association value. Step 2908
Then, from S1 = {"Computer", "Word Processor"}, S2
= {"Computer", "Outline font"}, the association value from S1 '= {"Calculator", "Word Processor"}, S2' = {"Calculator", "Outline font"}
Associative value to (400).

Further, in the specific example 7, in the search target comparison unit 104 of the specific example 3, the association value from a certain keyword to a certain keyword is changed from the synonym of the former keyword to the synonym of the latter keyword. It is also possible to calculate the priority.

<Effect> As described above, according to the specific example 7, the set S1 ′ of keywords obtained by replacing some of the input keywords from the set S1 of input keywords with synonyms.
From the set S2 of additional keywords, one or more sets S2 'of keywords obtained by replacing some of the additional keywords with synonyms are obtained, and the association value from S1 to S2 is S1' Since it is configured to obtain as an associative value to the
If they store the same number of associative values, a more appropriate priority can be given, and the effect that the searcher can not output the unintended search target in the search formula is further enhanced. can do.

<Specific Example 8><Structure> The structure of Specific Example 8 is the same except for the internal structure of the search target comparison unit 104 in Specific Example 1. The search target comparison unit 104 of the specific example 8 uses the synonym dictionary to set a keyword set S1 ′ in which some of the input keywords are replaced with synonyms, and an additional keyword from the additional keyword set S2. Among them, a set of keywords S2 'with some synonyms replaced is obtained so that the synonym of each keyword before and after the replacement (comprehensive synonym) is as large as possible, and the association value from S1 to S2 is S1. Obtained from the association value from'to S2 'and the synonym of the keywords before and after replacement,
The search target intended by the searcher is given a higher numerical value to the search target included in the set represented by the search formula.

FIG. 31 is an internal block diagram of the search target comparison unit in the eighth specific example. In the figure, 3101 obtains an association value from a set of input keywords to a set of additional keywords based on the association value in the association value synthesizing unit 3102, calculates the priority from this association value, and expresses it in a search formula. A priority addition unit that gives a higher priority to the search target included in the set as the search target intended by the searcher, 31
02 is a set of keywords S1 ′ in which some of the input keywords are replaced with synonyms based on the associative dictionary 3103 and the synonym dictionary 3104.
And a set of keywords S2 ′ obtained by replacing some of the additional keywords with synonyms from the set of additional keywords
Obtain the degree of synonym of each keyword before and after replacement as large as possible, and replace the association value from the input keyword set S1 to the additional keyword set S2 with the association value from the keyword set S1 'to the keyword set S2'. The associative value synthesis unit 3103, which is obtained from the similarities of the keywords before and after, is an associative dictionary that stores the associative value from a set of certain keywords to the set of certain keywords, and 3104 is a synonym dictionary that stores the similarities of the keywords. is there.

FIG. 32 is a diagram showing a specific example of the synonym dictionary 3204. As shown in the figure, for example, the synonym of “computer” and “information processing system” is 0.9.

<Operation> The operation of the specific example 8 is different only in the all associative value search processing of step 502 in the operation of the specific example 1.

FIG. 33 is a flowchart of the all-associative-value search process of the eighth specific example. First, in step 3301, the iterative process that continues until step 3304 is performed while lowering the overall degree of similarity to a certain threshold value. The overall degree of similarity is calculated by a method of multiplying all the degrees of similarity if the range of values of the degree of similarity between keywords before and after replacement is 0 or more and 1 or less. In step 3302, the synonym dictionary 3104 is accessed to obtain S1 to S1 ′ and S2 to S2 ′ so that the overall synonym degree is close to the value at that time. Where first S1 '
Get S1 when get, and get S2 when get S2 'first. In step 3303, the associative dictionary 3103 is S
If you remember the association value from 1'to S2 ', step 3
Proceed to 306, and if not stored, proceed to step 3304. In step 3305, the association value from S1 to S2 is set to zero. In step 3306, the associative value from S1 to S2 is a value calculated from the associative value from S1 ′ to S2 ′ and the similarity of the keywords before and after the replacement. This calculation is performed, for example, by setting the association value from S1 to S2 to a value obtained by multiplying the association value from S1 ′ to S2 ′ by the overall similarity.

An example of the operation of Concrete Example 8 will be briefly described below. FIG. 34 is an example of the association value stored in the association dictionary 3103. Assume that the association value from S1 = {“computer”, “word processor”} to S2 = {“computer”, “outline font”} is obtained. 1 of the repeated processing from step 3301 to step 3304
In the second round, the overall synonym is 1, S1 'is S1, S
You get S2 as 2 '. The associative dictionary 3103 does not store the associative value from S1 to S2. In the second iteration, the overall synonym is 0.8 and the overall synonym is 0.8.
As S1 ′ which is 1 (= 0.9 × 0.9), {“information processing system”, “word processor”} is obtained as S2 ′, and “{information processing system”, “outline font”} is obtained. In step 3306, the association value from S1 to S2 is set to 35
It is assumed that 0 (association value from S1 'to S2') x 0.81 (total degree of synonym) = 283.5.

In addition, in the specific example 8, in the search target comparison unit of the specific example 3, the synonym K1 ′ of the keyword K1 and the synonym K2 ′ of the keyword K2 are added to the synonym of each keyword (general synonym). It is also feasible to calculate the priority by obtaining the associativity from K1 to K2 from the associativity from K1 ′ to K2 ′ and the degree of similarity of the keywords.

Further, from the synonym dictionary 3104, not only the synonyms of the input keyword and the additional keyword but also the synonyms of the input keyword and the additional keyword can be obtained.

<Effect> As described above, according to the eighth specific example, a set S1 ′ of keywords in which some of the input keywords are replaced with synonyms from the set S1 ′ of input keywords.
And a set S2 ′ of keywords obtained by replacing some of the additional keywords with synonyms from the set S2 ′ of additional keywords.
Is the synonym of each keyword before and after replacement (general synonym)
Is as large as possible, and the association value from S1 to S2 is S
Since the associative value from 1'to S2 'and the synonym of the keywords before and after the replacement are obtained, the associative dictionary 3103 stores the same number of associative values as compared with the case (specific example 1) in which it is not. In this case, it is possible to give a more appropriate priority, and it is possible to further enhance the effect of preventing the searcher from outputting an unintended search target in the search expression.

<Specific Example 9><Structure> The structure of the specific example 9 includes a set of keywords obtained by replacing some of the input keywords (first set of keywords) S1 with synonyms (first set). 1 set of replacement keywords) S1 ′ and additional keyword set (second keyword set) S2, and a set of keywords obtained by replacing some of the additional keywords with synonyms (second set of replacement keywords) S2 'is obtained so that the synonyms (comprehensive synonyms) of the respective keywords before and after the substitution are as large as possible, and further, some of the keywords of S1' are replaced with synonyms (the third set). S1 "of the input keyword set S1", and a set of keywords obtained by replacing some of the keywords of S2 'with synonyms (fourth set of keywords) S2 ". The association value to a word from set S2, is obtained so as to obtain "the set S2 of the fourth replacement keyword" set S1 of the third substituted keywords calculated from association value to. That is, the specific example 9 is the search target comparison unit 10 in the specific example 8.
4 has a configuration in which an associative value complementing unit and a synonym dictionary are added.

FIG. 35 is a block diagram of the search object comparison unit in the ninth specific example. In the figure, 3501 is an associative value synthesizing unit 3
Based on the association value in 502 and the association value complementing unit 3503, the association value from the set of input keywords to the set of additional keywords is obtained, the priority is calculated from this association value, and the set is expressed by the search formula. For the included search target,
Based on the associative dictionary 3504 and the synonym dictionary 3506, the priority adding unit 3502 that gives a higher priority to a search target intended by the searcher makes some of the keywords into synonyms from the first keyword set. The first set of replacement keywords and the second set of replacement keywords obtained by replacing some of the second keywords with synonyms from the second set of keywords have as large a synonym as possible before and after replacement. Thus, the association value from the first set of keywords to the second set of keywords, the association value from the first set of replacement keywords to the second set of replacement keywords, and the synonyms of the keywords before and after replacement are obtained. The associative value synthesizing unit 3503, which is obtained from the degree, uses the third replacement keyword by replacing some of the keywords with synonyms. A set of replacement keywords and one or more sets of fourth replacement keywords obtained by replacing some of the keywords with synonyms from the second set of replacement keywords are obtained, and the second set of keywords is obtained from the first set of keywords. The associative value to the set of 3 is set as the associative value from the set of the third replacement keyword to the set of the fourth replacement keyword, and 3504 is the associative value from the set of certain keywords to the set of certain keywords. Associative dictionary to memorize, 3505
Is a synonym dictionary that stores a list of keywords that are synonyms, and 3506 is a synonym dictionary that stores the degree of synonym of keywords.

<Operation> The operation of the specific example 9 differs from the operation of the specific example 8 only in the operation of step 3303 in the specific example 8. In the specific example 9, step 3303.
Then, after replacing the set S1 ′ of input keywords, the associative value for the set S2 ′ of additional keywords is obtained by the all-associative-value search process of the seventh specific example. If the association value is not zero, the procedure proceeds to step 3306, and if it is zero, the procedure proceeds to step 3304.

<Effect> As described above, according to the specific example 9, a set S1 ′ of keywords obtained by replacing some of the input keywords from the set S1 of input keywords with synonyms.
From the set S2 of additional keywords, a set S2 'of keywords obtained by replacing some of the additional keywords with synonyms.
Is the synonym of each keyword before and after replacement (general synonym)
Is obtained as large as possible, and further, a set of keywords S1 'in which some of the keywords are replaced with synonyms S
1 ”and S2 'are obtained by replacing some of the keywords in S2' with synonyms S2", and the association value from S1 to S2 is calculated from the association value from S1 "to S2" Since it was done, compared with the case (specific example 7) not doing so,
If the associative dictionary 3504 stores the same number of associative values, it is possible to give a more appropriate priority and to prevent the searcher from outputting an unintended search target in the search formula. Can be large.

<Specific Example 10><Structure> In the structure of the specific example 10, the intention is clear when the search expression input by the searcher is an ambiguous search expression with respect to the structure of the specific example 1. This is a configuration in which a search formula changing unit for converting into a search formula is added.

FIG. 36 is a block diagram of the tenth specific example. In the figure, reference numeral 3602 denotes a search expression changing unit, and this search expression changing unit 3602 has a function of converting a search expression input by a searcher into a search expression having a clear intention when the search expression is ambiguous. I have. Also, this search formula changing unit 3602
As a method of converting the search expression input by the searcher into a search expression in which the searcher's intention is more clearly expressed, in, for example, Hiroshi Maruyama, Hideo Watanabe “Approximation of closeness of meaning between words by bit vector” Law ”The method shown in the 36th National Conference of Information Processing Society of Japan (the first half of 1988) 3U-9 etc. can be used. This method is a method of clarifying an intention by replacing a keyword whose searcher's intention is ambiguous with a keyword whose meaning distance is close to the keyword and whose intention is clear. Other components, the input unit 3
601 and search execution unit 3603 to output unit 3606 have the same configurations as the input unit 101 and search execution unit 102 to output unit 105 in the first specific example.

<Operation> The operation of the specific example 10 is different from the operation of the specific example 1 only in that in step 401 in the specific example 1, the search formula is changed immediately before searching the database.

FIG. 37 is a flowchart of the tenth specific example. That is, steps 3702 to 3704 in the flowchart shown are the same as steps 401 to 403 in the flowchart of the specific example 1 shown in FIG. Further, step 3701 is the above-mentioned search formula changing unit 3
Reference numeral 602 denotes a process of converting a search expression input by the searcher into a search expression having a clear intention when the search expression has an ambiguous intention.

Further, in the above-mentioned specific example 10, in order to give priority to the individual search targets included in the set by the search expression having a clear intention, the set of input keywords included in the search expression having an ambiguous intention is set. Instead, it is possible to give a more appropriate priority by using a set of input keywords included in the search formula whose intention is output by the search formula changing unit 3602. In the search target comparison unit 3605, it is also possible to use not a set of input keywords included in a search expression with an ambiguous intention, but a set of input keywords included in a search expression with a clear intention output by the search expression changing unit 3602. .

Further, in order to output the search target intended by the searcher without omission, the search expression changing unit 3602 outputs one or more search expressions having a clear intention, and these search expressions are expressed. It is also possible to give priority to the search targets included in the set.

<Effect> As described above, according to the tenth example, when the search expression input by the searcher is an ambiguous search expression, the search expression is converted into a clear search expression. Therefore, as compared with the case (specific example 1 to specific example 9) not doing so, the effect that the searcher does not output the search target that was not intended by the search expression, and the search that the searcher intended by the search expression The effect that the target can be output can be obtained.

<Specific Example 11><Structure> The structure of the specific example 11 differs from the structure of the specific example 10 described above in that a search target included in a set expressed by a search expression having a clear intention is Comprehensive priority calculation that gives priority (overall priority) that reflects both the priority of the converted search expression output by the search expression changing unit and the priority of the search target output by the search target comparison unit This is a configuration with additional parts.

FIG. 38 is a block diagram of the eleventh example. In the figure, reference numeral 3806 denotes the priority of the converted search expression output by the search expression changing unit 3802, and the search target comparison unit 3805.
Is a comprehensive priority calculation unit that assigns a comprehensive priority that reflects both the priorities for the search targets output by. Incidentally, the input unit 3801 to the search target comparison unit 3805, the output unit 380
7 is the same as the input unit 3601 to the search target comparison unit 3605 and the output unit 3606 in the tenth example.

<Operation> The operation of the concrete example 11 is the same as that of the concrete example 10.
The difference is that the total priority is calculated immediately after the search target comparison processing in step 3704 and the search target is output in accordance with the total priority.

FIG. 39 is a flowchart of the eleventh example. Steps 3901 to 3903 are the same as steps 3701 to 3703 in the tenth specific example. Also,
The calculation of the overall priority in step 3904 is performed by, for example, a method in which the overall priority is a value obtained by adding the priority of the converted search expression and the priority of the search target. Then, in step 3905, step 390
The search target based on the overall priority obtained in 4 is output.

In the eleventh specific example, the search expression changing unit 3802 outputs one or more combinations of search expressions with clear intentions and their priorities, and the overall priority calculating unit 3806.
However, it is also possible to give an overall priority to the search targets included in the set expressed by those search expressions.

<Effect> As described above, according to the eleventh example, the post-conversion converted by the search expression changing unit 3802 is output to the search target included in the set expressed by the search expression having a clear intention. Since the priority (overall priority) that reflects both the priority of the search formula and the priority of the search target output by the search target comparison unit 3805 is added, in the case of not doing so (specific example 1 to specific example 10) ), It is possible to obtain the effect that the searcher can not output the search target that was not intended by the search formula and that the searcher can output the search target that was intended by the search formula. it can.

<Specific Example 12><Structure> The structure of the specific example 12 differs from the structure of the specific example 1 in that the search target included in the set expressed by the search formula is
This is a configuration in which a search result division unit that divides into one or more sets is added.

FIG. 40 is a block diagram of the twelfth example. In the figure, 4004 is the search result dividing unit. Further, the input unit 4001 to the database 4003, the search target comparison unit 4005, and the output unit 4006 are the same as the input unit 101 to the database 103, the search target comparison unit 104, and the output unit 105 in the first specific example.

<Operation> FIG. 41 is a flowchart of the specific example 12. First, in step 4101, a database search is performed to obtain a search target included in the set represented by the search expression. Next, in step 4102, the search target included in the set expressed by the search expression is
Divide into one or more sets. The division of the search targets into one or more sets is performed by, for example, determining the number of sets according to the number of search targets and randomly selecting the search targets included in each set. In step 4103, the iterative process that continues until step 4106 in which the sets to be searched are used one by one is started. In step 4104, a search target comparison process is executed on the set of search targets being processed. In step 4105, the search target is output according to the priority.

<Effect> As described above, according to the specific example 12, the search target included in the set expressed by the search expression is
It is divided into one or more sets, the search target comparison process is executed for each search target set, and the search targets are output according to the priority. As a result, it is possible to obtain the effect that the time until the device starts responding to the searcher can be shortened.

<Specific Example 13><Structure> The structure of the specific example 13 differs from the structure of the specific example 12 in that search target comparison processing is performed for each set of one or more search targets divided by the search result dividing unit. This is a configuration in which a processing distribution unit assigned to one or more search target comparison units and one or more search target comparison units operating in parallel are added.

FIG. 42 is a block diagram of the embodiment 13. In the figure, reference numeral 4205 denotes a processing distribution unit that distributes search target comparison processing for each search target set divided by the search result division unit 4204, and reference numerals 4206-1 to 4206-N denote processing distribution units 420.
The search object comparison processing distributed in 5 is executed in parallel, and the searcher is included in the set represented by the search expression using the association value from the input keyword to the additional keyword. Is a search target comparison unit that gives a higher priority to a search target intended by. It should be noted that the input unit 4201 and the search result dividing unit 4204, which are other configurations, are used.
The output unit 4207 is the input unit 4 in the specific example 12.
001 to search result division unit 4004 and output unit 4006
This is the same configuration as.

<Operation> The operation of the specific example 13 differs from the operation of the specific example 12 in that step 410 in the specific example 12 is different.
Immediately after dividing the search result in step 2, the search target comparison processing for each of one or more sets of search targets is assigned to one or more search target comparison units 4206-1 to 4206-N, and from step 4103. The iterative process up to step 4106 is repeated for each of the search target comparison units 4206-1 to 4206.
The difference is that -N operates in parallel.

FIG. 43 is a flowchart of the specific example 13. That is, in steps 4301 to 4302,
After searching the database 4203, the search result dividing unit 4
204 divides the search results. And step 430
3, the processing distribution unit 4205 causes the search target comparison unit 42
The process is distributed to 06-1 to 4206-N, and in step 4304, the iterative process that continues until step 4307 in which the set to be searched is used one by one is started. In step 4305, a search target comparison process is executed for the set of search targets being processed. In step 4306, the search target is output according to the priority.

<Effect> As described above, according to the thirteenth specific example, one or more search object comparison units perform search object comparison processing for each of one or more sets of search objects divided by the search result dividing unit 4204. No. 4206-1 to 4206-N, and the search target comparison processing is executed in parallel. Therefore, as compared with the case (specific example 1 to specific example 12) which does not do so, the apparatus ends the response to the searcher. The effect that the time can be shortened can be obtained.

In the specific examples 1 to 13, the retrieval target is a document, but the present invention is not limited to this and can be applied to various retrieval targets such as voice and images. Further, in these specific examples, the information search device in which the searcher inputs a search expression has been described, but the present invention is also applicable to use as a part of an information processing device that performs a database search process in some form.

[Brief description of the drawings]

FIG. 1 is a configuration diagram of a specific example 1 in an information search device of the present invention.

FIG. 2 is a configuration diagram of a search target comparison unit in specific example 1 of the information search device of the invention.

FIG. 3 is an explanatory diagram of an associative dictionary in a specific example 1 of the information search device of the invention.

FIG. 4 is a flowchart of specific example 1 in the information search device of the present invention.

FIG. 5 is a flowchart of a search target comparison process of specific example 1 in the information search device of the invention.

FIG. 6 is a flowchart of an all-associative-value search process of specific example 1 in the information search device of the invention.

FIG. 7 is an explanatory diagram of an association value of a specific example 1 in the information search device of the invention.

FIG. 8 is a configuration diagram of a search target comparison unit in a specific example 2 of the information search device of the invention.

FIG. 9 is a flowchart of a search target process of the specific example 1 in the information search device of the invention.

FIG. 10 is a configuration diagram of a search target comparison unit in specific example 3 of the information search device of the invention.

FIG. 11 is an explanatory diagram of an associative dictionary in Specific Example 3 of the information search device of the invention.

FIG. 12 is a flowchart of an all-associative-value search process in a specific example 3 of the information search device of the invention.

FIG. 13 is a flowchart of each associative value search process in the third specific example of the information search device of the present invention.

FIG. 14 is an explanatory diagram of an association value in a specific example 3 of the information search device of the invention.

FIG. 15 is a configuration diagram of a search target comparison unit in specific example 4 of the information search device of the invention.

FIG. 16 is a flowchart of search target comparison processing of specific example 4 in the information search device of the present invention.

FIG. 17 is a flowchart of an all-associative value search process of specific example 4 in the information search device of the invention.

FIG. 18 is a flowchart of each associative value search process of specific example 4 in the information search device of the invention.

FIG. 19 is an explanatory diagram of an association value in Example 4 of the information search device of the invention.

FIG. 20 is an explanatory diagram of an association value obtained in step 1603 of specific example 4 of the information search device of the invention.

FIG. 21 is a configuration diagram of a search target comparison unit in specific example 5 of the information search device of the invention.

FIG. 22 is a flowchart of each associative value search process in specific example 5 of the information search device of the invention.

FIG. 23 is an explanatory diagram of an association value in a specific example 5 of the information search device of the invention.

FIG. 24 is a diagram showing a directed graph in specific example 5 of the information search device of the invention.

FIG. 25 is a diagram showing a directed graph in a specific example 6 of the information search device of the invention.

FIG. 26 is a diagram showing a directed graph in a modified example of the sixth specific example of the information search device of the present invention.

FIG. 27 is a configuration diagram of a search target comparison unit in Specific Example 7 of the information search device of the invention.

FIG. 28 is an explanatory diagram of a synonym dictionary in a specific example 7 of the information search device of the invention.

FIG. 29 is a flowchart of an all-association value search process in the specific example 7 of the information search device of the invention.

FIG. 30 is an explanatory diagram of an association value in a specific example 7 of the information search device of the invention.

FIG. 31 is a configuration diagram of a search target comparison unit in Specific Example 8 of the information search device of the invention.

FIG. 32 is an explanatory diagram of a synonym dictionary in a specific example 8 of the information search device of the invention.

FIG. 33 is a flowchart of an all-association value search process in the specific example 8 of the information search device of the invention.

FIG. 34 is an explanatory diagram of an association value in a specific example 8 of the information search device of the invention.

[Fig. 35] Fig. 35 is a configuration diagram of a search target comparison unit in Specific Example 9 of the information search device of the present invention.

FIG. 36 is a configuration diagram of a specific example 10 in the information search device of the invention.

FIG. 37 is a processing flowchart of specific example 10 in the information search device of the invention.

FIG. 38 is a configuration diagram of a specific example 11 in the information search device of the invention.

FIG. 39 is a processing flowchart of specific example 11 in the information search device of the invention.

FIG. 40 is a configuration diagram of a specific example 12 in the information search device of the invention.

FIG. 41 is a processing flowchart of specific example 12 in the information search device of the invention.

FIG. 42 is a configuration diagram of a specific example 13 in the information search device of the invention.

FIG. 43 is a processing flowchart of specific example 13 in the information search device of the invention.

[Explanation of symbols]

101, 3601, 3801, 4001, 4201 Input unit 102, 3603, 3803, 4002, 4202 Search execution unit 103, 3604, 3804, 4003, 4203 Database 104, 3605, 3805, 4005, 4206-1
~ N Search target comparison unit 105, 3606, 3807, 4006, 4207 Output unit 201, 801, 1001, 1501, 2101, 27
01, 3101, 3501 Priority addition section 202, 803, 1003, 1504, 2104, 27
03 3103 3504 Associative dictionary 802 Keyword selection section 1002, 1502, 2102 Priority calculation section 1503 Keyword set calculation section 2103 Rule connection section 2702, 3503 Associative value complementing section 2704, 3505 Synonym dictionary 3102, 3502 Associative value composition section 3104 , 3506 Synonym dictionary 3602, 3802 Search expression changing unit 3806 Overall priority calculating unit 4004, 4204 Search result dividing unit 4205 Processing distribution unit

Claims (21)

[Claims] 1. An input unit for inputting a search expression, a database for storing a search target, a search for the search target of the database based on the search expression input from the input unit, and the search. A search execution unit that outputs a search target included in the set represented by the expression, and an association value from the set of input keywords included in the search formula to the set of additional keywords added to the search target. , The search target comparison unit that gives a higher priority to the search target that the searcher intends with respect to the search target included in the set represented by the search expression, and the search execution unit output An information retrieving apparatus, comprising: an output unit that outputs a retrieval target according to the priority given to the retrieval target by the retrieval target comparison unit. 2. The information search device according to claim 1, wherein the search target comparison unit stores an associative dictionary for storing an association value from a set of keywords to a set of keywords, and a set from a set of input keywords to additional keywords. The associativity value is calculated using the associative dictionary, and the associativity value is higher in the priority of the search target intended by the searcher than the search target included in the set expressed by the search expression. An information retrieving apparatus, comprising: 3. An input unit for inputting a search expression, a database for storing a search target, a search for the search target in the database based on the search expression input from the input unit, and the search. A search execution unit that outputs a search target included in the set expressed by the expression, and a keyword selection unit that selects a set of input keywords and a set of additional keywords that do not affect the result of the search by the search execution unit. , An associative dictionary that stores an associative value from a certain keyword set to a certain keyword set, and an associative value from an input keyword set selected by the keyword selection unit to an additional keyword set, using the associative dictionary The priority that is higher than the search target intended by the searcher with respect to the search target included in the set expressed by the search expression. A search target comparison unit having a priority addition unit attached to the search target, and an output unit that outputs the search target according to the priority assigned to the search target comparison unit with respect to the search target output by the search execution unit. An information retrieval device characterized by being provided. 4. An input unit for inputting a search formula, a database for storing a search target, a search for the search target of the database based on the search formula input from the input unit, and the search. A search execution unit that outputs the search target included in the set expressed by the expression, an associative dictionary that stores the association value of a keyword to a keyword, and an additional keyword from the input keyword based on the association value of the association dictionary To the search target included in the set expressed by the search formula, and the priority calculated by the priority calculation unit. On the basis of the search target output unit, the search target comparison unit includes a priority addition unit that assigns a higher priority to a search target intended by the searcher, and the search target output from the search execution unit. Information retrieval apparatus characterized by comprising an output section for outputting the search according to the priority granted in elephants comparison unit. 5. The information retrieval apparatus according to claim 4, wherein an association dictionary that stores an association value from a keyword to a keyword, a cache memory that stores the association value read from the association dictionary, and the association Based on the associative value stored in the dictionary or the associative value stored in the cache memory, the associative value from the input keyword to the additional keyword is obtained, and the priority calculation unit that calculates the priority from this associative value and the search formula are Based on the priority obtained by the priority calculation unit, for the search target included in the represented set,
An information retrieval apparatus comprising: a retrieval target comparison unit including a priority addition unit that gives a higher priority to a retrieval target intended by a searcher. 6. An input unit for inputting a search formula, a database for storing a search target, a search for the search target in the database based on the search formula input from the input unit, and the search. A search execution unit that outputs the search target included in the set represented by the expression, an associative dictionary that stores an association value from a certain keyword set to a certain keyword set, and a keyword set that calculates a subset of the certain set A calculation unit and an association value from the subset of the set of input keywords calculated by the keyword set calculation unit to the subset of the set of additional keywords are obtained based on the association dictionary, and the priority is calculated from this association value. And a search target that is included in the set expressed by the search formula, based on the priority calculated by the priority calculation unit. The search target comparison unit having a priority addition unit that gives a higher priority to the elephant, and the search target output from the search execution unit, according to the priority given by the search target comparison unit And an output unit for outputting the information retrieval device. 7. An input unit for inputting a search formula, a database for storing a search target, a search for the search target of the database based on the search formula input from the input unit, and the search. A search execution unit that outputs the search target included in the set expressed by the expression, an associative dictionary that stores the association value from a certain keyword set to a certain keyword set, and another set that is different from a certain set Based on the associative dictionary, an association value from another set different from the set of input keywords calculated by the keyword set calculation unit to another set different from the set of additional keywords is calculated based on the association dictionary. , Based on the priority calculated by the priority calculation unit for the search target included in the set expressed by the search formula, and the priority calculation unit that calculates the priority from the association value A search target comparison unit having a priority addition unit that assigns a higher priority to a search target intended by the searcher, and a search target comparison unit for the search target output by the search execution unit. An information retrieval apparatus comprising: an output unit that outputs a retrieval target according to the assigned priority. 8. An input unit for inputting a search formula, a database for storing a search target, a search for the search target of the database based on the search formula input from the input unit, and the search. The search execution unit that outputs the search target included in the set expressed by the expression, the association dictionary that stores the association value from a keyword to a keyword, and the degree to which a searcher associates a keyword with a keyword From the input keyword to the additional keyword based on the rule connection part that calculates the association value by connecting one or more rules that are association values and the association value of the association dictionary and the connection result in the rule connection part. And the priority calculation unit for the search target included in the set represented by the search formula. Based on the priority obtained in, the search target comparison unit having a priority addition unit for giving a higher priority to the search target intended by the searcher, and the search target output by the search execution unit And an output unit that outputs a search target according to the priority given by the search target comparison unit. 9. The information search device according to claim 8, wherein the searcher links one or more rules that a degree of associating a certain keyword out of a certain keyword is the associative value with a number equal to or less than a certain threshold, An information retrieval device comprising a rule connecting unit for calculating an association value from a directed graph obtained first. 10. The information search device according to claim 8, wherein the searcher links one or more rules that a degree of associating a certain keyword out of a certain keyword is the associative value, and the rule is less than a certain threshold value. An information retrieval device comprising a rule connecting unit for calculating an associative value from each associative value calculated from one or more directed graphs. 11. The information retrieval apparatus according to claim 8, wherein the associative dictionary stores at least one rule that a degree that a searcher associates a certain keyword with a certain keyword is its associated value. An information retrieving apparatus, characterized by comprising a rule connecting section for calculating association values from respective association values calculated from one or more directed graphs until all rules having the same existing node as a starting point are connected. 12. An input unit for inputting a search formula, a database for storing a search target, a search for the search target of the database based on the search formula input from the input unit, and the search. A search execution unit that outputs a search target included in the set represented by the expression, an associative dictionary that stores an association value from a keyword to a keyword, and a keyword set calculation unit that calculates a subset of the certain set, A rule connecting unit that calculates an association value by connecting one or more rules that a searcher associates a keyword with a keyword with the association value, an association value of the association dictionary, and the keyword set. From the subset of the input keyword set to the subset of the additional keyword set based on the subset obtained by the calculation unit and the connection result by the rule connection unit Based on the priority calculated by the priority calculation unit for the search target included in the set expressed by the search formula and the priority calculation unit that calculates the association value of the A search target comparison unit having a priority addition unit that assigns a higher priority to a search target intended by the searcher, and a search target comparison unit for the search target output by the search execution unit. An information retrieval apparatus comprising: an output unit that outputs a retrieval target according to the assigned priority. 13. An input unit for inputting a search formula, a database for storing a search target, a search for the search target of the database based on the search formula input from the input unit, and the search. A search execution unit that outputs a search target included in the set represented by the expression, an associative dictionary that stores an association value from a keyword to a keyword, and a keyword set calculation unit that calculates a subset of the certain set, One or more rules that a searcher associates a certain keyword with a certain keyword is the associative value, and the rule used for the connection is a set of one keyword to a set of the other keyword. Rule association part for calculating the association value by calculating from the association value for the subset of, and the association value of the association dictionary and the keyword set calculation. The association value from the subset of the set of input keywords to the subset of the set of additional keywords is calculated based on the subset obtained by the section and the result of connection by the rule connection unit, and the priority is calculated from this association value. Based on the priority calculated by the priority calculation unit, the priority calculation unit and the search target included in the set represented by the search formula have a higher numerical value as the search target intended by the searcher. A search target comparison unit having a priority addition unit for adding a search target, and an output unit that outputs the search target according to the priority given by the search target comparison unit with respect to the search target output by the search execution unit. An information retrieval device characterized in that 14. An input unit for inputting a search formula, a database for storing a search target, a search for the search target of the database based on the search formula input from the input unit, and the search. A search execution unit that outputs the search target included in the set represented by the expression, an associative dictionary that stores an association value from a keyword to a keyword, and a synonym dictionary that stores a list of synonymous keywords. , Based on the associative dictionary and the synonym dictionary, obtain one or more set of keywords S1 'by replacing some of the input keywords with synonyms from the set of input keywords S1 and obtain one or more of the additional keywords from the set of additional keywords S2 A set of keywords with some synonyms replaced
S1 'or more is obtained, and the association value from the set S1 of input keywords to the set S2 of additional keywords is set to the set S1' of keywords.
Based on the association value complementing section as an association value from the keyword set S2 ', and the association value in the association value complementing section, obtain the association value from the set of input keywords to the set of additional keywords, and from this association value Search target comparison that has a priority addition unit that calculates a priority and assigns a higher priority to a search target included in the set expressed by the search formula as the search target intended by the searcher An information retrieval apparatus comprising: a unit and an output unit that outputs a search target to the search target output by the search execution unit according to the priority given by the search target comparison unit. 15. An input unit for inputting a search formula, a database for storing a search target, a search for the search target of the database based on the search formula input from the input unit, and the search. A search execution unit that outputs a search target included in a set represented by an expression, an associative dictionary that stores an association value from a keyword to a keyword, a synonym dictionary that stores a degree of similarity of the keyword, and the association. Based on the dictionary and the synonym dictionary, a set of keywords S1 'obtained by replacing some of the input keywords with synonyms from the set of input keywords S1' and a set of keywords obtained by replacing some of the additional keywords with synonyms from the set of additional keywords S2 The set S2 'is obtained so that the similarity of each keyword before and after replacement is as large as possible, and the set of additional keywords is obtained from the set S1 of input keywords. Associative value to the association S2, the association value from the set S1 'of keywords to the set S2' of keywords and the associativity synthesizing unit which is obtained from the similarity of the keywords before and after the replacement, Based on
The search that the searcher intends for the search target included in the set expressed by the search formula by calculating the association value from the set of input keywords to the set of additional keywords, calculating the priority from this The search target comparison unit having a priority addition unit that assigns a higher priority to the target, and the search target output by the search execution unit, according to the priority given by the search target comparison unit And an output unit for outputting the information retrieval device. 16. An input unit for inputting a search formula, a database for storing a search target, a search for the search target of the database based on the search formula input from the input unit, and the search. A search execution unit that outputs the search target included in the set represented by the expression, an associative dictionary that stores an association value from a keyword to a keyword, and a synonym dictionary that stores a list of synonymous keywords. A synonym dictionary that stores synonym degrees of keywords, and a first set of replacement keywords in which some of the first keywords are replaced with synonyms based on the associative dictionary and the synonym dictionary,
A second set of replacement keywords obtained by replacing some of the keywords with synonyms from the second set of keywords is obtained so that the degree of synonym of each keyword before and after replacement is as large as possible. An associative value synthesizer for obtaining an associative value for the second set of keywords from the associative value from the first set of replacement keywords to the second set of replacement keywords and the similarity of the keywords before and after the replacement; A third set of replacement keywords in which some of the replacement keywords are replaced with synonyms,
One or more fourth replacement keyword sets obtained by replacing some of the keywords with synonyms are obtained from the second replacement keyword set, and the association value from the first keyword set to the second keyword set is obtained. Is an association value from the third set of replacement keywords to the fourth set of replacement keywords, and the set of additional keywords from the set of input keywords based on the association value in the association value complementing unit. To the search target included in the set represented by the search formula,
A search target comparison unit having a priority addition unit that assigns a higher priority to the search target intended by the searcher, and the search target comparison unit assigns to the search target output from the search execution unit. And an output unit that outputs a search target according to the priority. 17. An input unit for inputting a search expression, a database for storing a search target, and a search expression input from the input unit, when the intention is ambiguous, the intention is clear. Based on the search formula output by the search formula change unit and the search formula change unit, the search target is included in the set represented by the search formula. A search execution unit that outputs a search target, and a set expressed by the search formula using an association value from a set of input keywords included in the search formula to a set of additional keywords added to the search target. The search target comparison unit that gives a higher priority to the search target that the searcher intends for the search target included in the search target comparison unit that outputs the search target output from the search execution unit. Granted in part Information retrieval apparatus characterized by comprising a an output unit that outputs the search target in accordance with the priority. 18. An input unit for inputting a search expression, a database for storing a search target, and a search expression input from the input unit, when the intention is ambiguous, the intention is clear. Based on the search formula output by the search formula change unit and the search formula change unit, the search target is included in the set represented by the search formula. A search execution unit that outputs a search target, and an association value from the input keyword to the additional keyword included in the search formula whose intention is output by the search formula changing unit is included in the set expressed by the search formula. The search target comparison unit that gives a higher priority to the search target that the searcher intends, and the search target comparison unit outputs the search target output by the search execution unit. Granted priority Information retrieval apparatus characterized by comprising an output section for outputting the searched according. 19. An input unit for inputting a search expression, a database for storing a search target, and a search expression input from the input unit, when the intention is ambiguous, the intention is clear. Based on the search formula output by the search formula change unit and the search formula change unit, the search target is included in the set represented by the search formula. A search execution unit that outputs a search target, and an association value from an input keyword to an additional keyword included in the search formula whose intention is output by the search formula changing unit is included in the set expressed by the search formula. A search target comparison unit that assigns a higher priority to a search target that the searcher intends, a priority for the converted search formula output by the search formula changing unit, and the search target The comparison section comes out Information retrieval apparatus, comprising: an overall priority calculation unit that assigns an overall priority that reflects both priorities for the retrieval target to be searched; and an output unit that outputs the retrieval target according to the overall priority. . 20. An input unit for inputting a search expression, a database for storing a search target, a search for the search target of the database based on the search expression input from the input unit, and the search. A search execution unit that outputs a search target included in the set represented by the expression; a search result division unit that divides the search target included in the set represented by the search expression into one or more sets; For each set of search targets divided by the search result division unit,
Using the association value from the set of input keywords included in the search formula to the set of additional keywords added to the search target, the search target included in the set represented by the search formula is searched. The search target intended by a person has a search target comparison unit that gives a higher priority to the search target, and an output unit that outputs the search target according to the priority assigned by the search target comparison unit. Search device. 21. An input unit for inputting a search formula, a database for storing a search target, a search for the search target of the database based on the search formula input from the input unit, and the search. A search execution unit that outputs the search target included in the set expressed by the expression, and a search result division unit that divides the search target included in the set expressed by the search expression into one or more search target sets. , A processing distribution unit that distributes a search target comparison process for each search target set divided by the search result division unit, and a search target comparison process that is distributed by the process distribution unit, respectively executed in parallel, and input keyword One or more searches that give higher priority to the search target that the searcher intends for the search target that is included in the set expressed by the search expression, using the association value from Information retrieval device comprising an elephant comparing unit, further comprising an output unit that outputs the search target in accordance with the priority of the search target comparison unit.