JP2012074087A - Document retrieval system, document retrieval program, and document retrieval method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a document retrieval method, a document retrieval system, and a document retrieval program capable of appropriately determining an order of precedence of documents output as a retrieval result, when document retrieval is performed by using an input keyword and a translated keyword.SOLUTION: Keyword translation means 22 of a document retrieval device 10 translates an input keyword to be a translation keyword. Keyword score determination means 23 determines a keyword score for the input keyword and the translation keyword respectively. Document retrieval means 24 retrieves documents on the basis of the input keyword and the translation keyword, and obtains a plurality of documents. Document score calculation means 25 calculates a document score for each of the documents of a retrieval result on the basis of the keyword score and appearance frequency of each keyword. Retrieval result output means 26 determines an order of the documents of the retrieval result in descending order of the document score and performs output according to the order.

Description

  The present invention relates to a document search system and a document search method for searching for a document using a keyword, and more particularly to a method using a translated keyword. The present invention also relates to a document search program that causes a computer to function as such a document search system.

  In a document search system, when a document database includes documents in a plurality of languages, a system that translates an input keyword into another language as a keyword used for search is known. An example of such a system is described in US Pat. Patent Literature 1 describes that a keyword specified in Japanese is translated into English, Japanese literature is searched using Japanese, and English literature is searched using English. ing.

Japanese Patent Laid-Open No. 10-232883

However, in the conventional technique, when a search is performed using a plurality of languages, there is a problem that the priority order of documents output as a search result cannot be determined appropriately.
In general, since words are ambiguous, an optimal selection is not always made when a keyword input in a native language is translated into another language. For this reason, when determining the priority order of the document in the document list of the search result, for example, the priority order may not be appropriately determined for the document including the translated keyword.

The present invention has been made to solve such problems, and when searching for a document using an input keyword and a translated keyword, the priority order of the document output as a search result is set. An object of the present invention is to provide a document search system and a document search method that can be appropriately determined.
A further object of the present invention is to provide a document search program that causes a computer to function as such a document search system.

A document search system according to the present invention is a document search system including a document search device that searches for a document using a keyword, a translation service device, and a document database that stores a plurality of documents to be searched. And
The document search apparatus includes a keyword receiving unit that receives one or more keywords as input keywords, and a keyword that acquires, for each of the input keywords, translation keywords obtained by translating the input keywords into other languages for a plurality of other languages. A translation means; a keyword score determination means for determining a keyword score for each of the input keyword and the translation keyword; a document search means for searching a document based on the input keyword and the translation keyword and obtaining a plurality of search result documents; and a search A translation service device comprising: a document score calculating unit that calculates a document score based on a keyword score for each result document; and a search result output unit that outputs each search result document in association with a corresponding document score. That of the input keyword Correspondingly, a plurality of translated keywords having a ranking are generated, the translation service device represents the ranking of the translated keywords based on the posting order of the translated keywords, and the keyword score determining means includes each of the input keywords, The translation score is determined based on the ranking for all combinations with each of the keywords, and the keyword score determining means determines the keyword score based on all of the related translation scores for each of the translation keywords, and the keyword of the input keyword The score is higher than any of the keyword scores of the translation keywords corresponding to the input keyword.

The document database stores language information indicating in which language the document is expressed in association with each document, and the translation service device indicates in which language the translation keyword is expressed for each translation keyword. And the translation service device stores language information representing a predetermined language corresponding to the input keyword, and the document score calculating means calculates the document score of the document. Of the input keyword and the translation keyword, the keyword score of the language information that matches the language information of the document may be applied.
The document search system accepts designation of whether or not to use the translation keyword, and the document score calculation means calculates the document score based on the input keyword and the translation keyword according to the designation, or inputs regardless of the translation keyword Whether to calculate the document score based only on the keyword may be switched.
The document score calculation means may further calculate the document score based on the number of times each of the input keyword and the translation keyword appears in the search result document.
The document score calculation means may further calculate the document score based on the recognition rate in the character recognition process for the search result document.
The document search means has a function of searching for a document by OR search and a function of searching for a document by AND search when the keyword receiving means receives two or more input keywords. In the AND search, the document search means searches for all the input keywords “for the input keyword and each translation keyword corresponding to the input keyword” in the AND search. A document that satisfies the condition that “at least one of them appears” may be searched.

  A document search program according to the present invention causes a computer to function as the above-described document search system.

  A document search method according to the present invention is a document search method in which the above-described document search system searches for a document.

  Since the document search method and the document search system according to the present invention determine the keyword score for each of the input keyword and the translated keyword and calculate the document score based on the keyword score, the document score is output as a search result. Document priorities can be appropriately determined.

These are figures which show the structure of the document search system based on this invention. 6 is a flowchart for explaining the operation of the document search apparatus in the document search system of FIG. 1. It is a figure which shows the example of the correspondence of an input keyword and a translation keyword. It is a figure which shows the example of the correspondence of the ranking of a translation keyword, and the translation score based on the ranking. It is a figure which shows the example of the correspondence of the translation score based on an order | rank about each keyword, and the keyword score finally provided with respect to each keyword. It is a figure which shows the example of the information showing the frequency | count that each keyword appears in the text data of a search result document. It is a figure which shows the example of the calculation result of the document score with respect to a search result document.

  In the present invention, when a keyword is input in a certain language in a search from a document database including documents written in various languages such as Japanese, English, French, and Chinese, the input keyword is converted into another language by a translation engine. A search is performed using the input keyword and the keyword converted into another language at the same time. Priorities among keywords are given by assigning scores to the keywords, and the priorities are reflected in the priorities of search results of documents and output. In this way, a document search method corresponding to multiple languages is realized.

Embodiments of the present invention will be described below with reference to the accompanying drawings.
Embodiment 1 FIG.
FIG. 1 shows a configuration of a document search system 100 according to the present invention. The document search system 100 is a system used to search for a document using a keyword.
The document search system 100 includes a document search apparatus 10 that searches for a document using a keyword.

The document search apparatus 10 is an information processing apparatus and has a configuration as a well-known computer.
The document search apparatus 10 includes an input device 30 that is used by a user to input keywords. The input device 30 is, for example, a mouse or a keyboard. In addition, the document search device 10 includes a display device 40 that displays a search processing result to the user. The display device 40 is, for example, a display or a printer. In addition, the document search device 10 includes a calculation device 20 that performs a calculation. The arithmetic unit 20 is, for example, a CPU (Central Processing Unit).
Although not shown, the document search apparatus 10 includes a memory and an HDD (hard disk drive) as storage means for storing information. In addition, the document search apparatus 10 includes a network interface for transmitting / receiving information to / from other information processing apparatuses.

The storage unit of the document search device 10 stores a document search program that defines the operations of the document search device 10 and the arithmetic device 20. The arithmetic unit 20 executes the document search program, thereby executing the keyword receiving unit 21, the keyword translating unit 22, the keyword score determining unit 23, the document searching unit 24, the document score calculating unit 25, and the search result output shown in FIG. Although it functions as the means 26, the detail of each function is mentioned later.
In addition, the document search device 10 that is a computer realizes other functions described in the present specification by causing the computing device 20 to execute a document search program or another program.

The document search system 100 includes a translation service device 110 that is communicably connected to the document search device 10. The translation service device 110 translates keywords. The translation service device 110 receives a phrase expressed in a certain language, translates it into another language, and outputs it. That is, it has a function of generating a keyword (translation keyword) obtained by translating the input keyword into another language based on the input keyword (input keyword). Note that “translation” here can also be understood as conversion from a keyword in one language to a keyword in another language.
The translation service device 110 performs translation into a plurality of languages. For example, for an input keyword in Japanese, a translation keyword in English and a translation keyword in French are generated and output.

In addition, the translation service device 110 generates a plurality of translation keywords having ranks for one input keyword. That is, for a certain word, for example, according to the frequency with which the corresponding translated word is used, a ranking is given in order from the most frequently used translated word to generate a list of translated keywords. This list represents, for example, the ranking of each translation keyword by arranging the translation keywords in accordance with the ranking, but the ranking of each translation keyword may be represented by associating the translation keyword with a numerical value representing the ranking.
A well-known configuration can be used for the translation service device 110. For example, the translation service device 110 stores a dictionary file that associates one or more translated words with a large number of words and phrases, and translates with reference to the dictionary file.

The document search system 100 includes a document database 120 that is communicably connected to the document search apparatus 10. The document database 120 stores a plurality of documents to be searched by the document search device 10.
The document database 120 receives input of one or more keywords, extracts all of the stored documents including any of the keywords, and outputs the extracted document or a list thereof.

The operation of the document search system 100 configured as described above will be described using the flowchart of FIG. 2 and the data examples of FIGS.
FIG. 2 is a flowchart for explaining the operation of the document search apparatus 10 in the document search system 100. First, the keyword receiving means 21 receives one or more input keywords used for the search from the user via the input device 30 (step S1, keyword receiving step). In this example, it is assumed that two Japanese input keywords “teacher” and “teacher” are received.

  Next, the keyword translation means 22 uses the translation service device 110 to translate the input keyword into a translation keyword (step S2, keyword translation step). In step S <b> 2, the keyword translation unit 22 passes the input keyword to the translation service device 110, and the translation service device 110 generates a translation keyword for each received input keyword and returns it to the keyword translation unit 22. Thus, the keyword translation means 22 acquires a translation keyword.

  FIG. 3 shows an example of the correspondence between input keywords and translation keywords. In this example, there are two types of translation keywords, one in English as shown in FIG. 3 (a) and one in French as shown in FIG. 3 (b). In the table of FIG. 3A, for the input keyword “teacher”, there are three English translation keywords “teacher” having rank 1, “instructor” having rank 2, and “master” having rank 3. Are associated. Thus, the translation service device 110 stores a plurality of ranked translation keywords in association with each of the input keywords.

In the table of FIG. 3B, the input keyword “teacher” is also associated with two French translation keywords “professeur” having rank 1 and “instructeur” having rank 2. . Thus, the keyword translation means 22 acquires a translation keyword about several languages other than the language of an input keyword.
Note that the document search apparatus 10 may store the input keyword, the acquired translation keyword, and the correspondence shown in FIG. 3 in a storage unit in the form of a table or the like.

  Next, the keyword score determination means 23 determines a keyword score for each of the input keyword and the translation keyword (step S3, keyword score determination step). Here, the keyword score determining means 23 determines the keyword score based on the correspondence relationship shown in FIGS.

FIG. 4 shows an example of the correspondence between the ranking of translation keywords and the translation score based on the ranking. The keyword score determination means 23 determines the keyword score of each translation keyword based on this translation score. The document search apparatus 10 stores in advance the correspondence shown in FIG. 4 in the form of a table or the like in its storage means, and this correspondence can be changed as appropriate by the user or administrator of the document search apparatus 10. Can do.
A predetermined score that is always constant, for example, 100 is given to the input keyword (note that this score is treated in part differently from the translation score, as will be described later, and is shown in parentheses in FIG. ). Also, different translation scores are given to the translation keywords depending on the ranking. The score to be given decreases by a predetermined value, for example, by 10 every time the rank is lowered, and becomes 90 for rank 1, 80 for rank 2, and 70 for rank 3.

A larger score value means that a document including the keyword is more important in the search result (that is, the document rank is higher in the search result). Thus, the keyword score determination means 23 determines the keyword score of a translation keyword based on the order of a translation keyword.
The relationship between the ranking and the translation score is not limited to that shown in FIG. The translation score for rank 1 may be a value lower than the keyword score for the input keyword. The translation score for rank 2 or lower may be a function that monotonously decreases as the rank decreases (that is, as the number representing the rank increases in this example).

The translation service device 110 ranks translation keywords according to the frequency generally used as a translation word. Here, when information such as syntax and context is not considered, a translation that is frequently used as a translation among a plurality of translations described in a dictionary or the like for a certain phrase may be a more appropriate translation. it can. A document that includes a more appropriate translation is more likely to be desired by the user than a document that includes only a translation that is not. That is, a translation keyword with a higher rank can be said to be a more reliable keyword. Since the keyword score determining means 23 determines the translation score based on the ranking of each translation keyword, the translation score of the more likely translation keyword can be increased, and a more accurate search result can be obtained.
Note that the translation service device 110 need not strictly rank each keyword according to the frequency of use by strictly statistical analysis. General dictionaries usually determine the order in which they are placed, taking into account the frequency of use of translated words, etc., so using a well-known dictionary as a general dictionary has some effect on improving the accuracy of search results. Can be obtained.

FIG. 5 shows an example of the correspondence relationship between the translation score based on the ranking and the keyword score finally given to each keyword for each keyword.
The keyword score determining means 23 always gives a keyword score of 100 to the input keyword as described above. For translation keywords, first, for all combinations of each input keyword and each translation keyword, a translation score is determined based on the ranking. In FIG. 5, the translation score based on the ranking is given for all combinations (two types in total) of two input keywords and five translation keywords.

  For example, as shown in FIG. 3A, the translation keyword “master” has a rank 3 for the input keyword “teacher”, and therefore a translation score 70 corresponding to the rank 3 is given in FIG. Also, since this translation keyword “master” has a ranking 2 for the input keyword “teacher”, a translation score 80 corresponding to the ranking 2 is given in FIG. When a certain translation keyword has no rank with respect to any input keyword, that is, when the translation keyword is not a translation of the input keyword, the translation score for the combination is set to 0. However, the translation score in this case may not be 0, and may be a value smaller than any of the translation scores of other translation keywords corresponding to the input keyword.

Based on the translation score thus determined, the keyword score determination means 23 further determines a final keyword score for each translation keyword. In the example of FIG. 5, the average of translation scores given to the translation keyword is used as the keyword score of the translation keyword.
Thus, the keyword score determination means 23 determines a keyword score based on all the related translation scores for each translation keyword.
Note that the document search apparatus 10 may store the correspondence shown in FIG. 5 in the storage unit in the form of a table or the like.

Here, the keyword score given to the input keyword is always 100 as described above. Further, since all translation scores are 90 or less (that is, translation scores for rank 1), the average keyword score (keyword score of translation keyword) is always 90 or less. Therefore, the keyword score given to the input keyword in the native language is higher than any of the keyword scores given to the translation keywords in other languages.
Since the input keyword in the native language has no fear of mistranslation or inappropriate translation, the document including the input keyword is more likely to be desired by the user than the document including only the translation keyword. That is, the input keyword can be said to be a more probable keyword. In this way, a more accurate search result can be obtained by setting a higher score for a more certain input keyword and placing importance on it, and setting a relatively low score for the translation keyword.

In addition, when a certain translation keyword corresponds to a plurality of input keywords, such as the translation keyword “master” in this example, a document including that translation keyword is more desirable by the user than a document including only other translation keywords. There is a high possibility that That is, such a translation keyword can be said to be a more reliable keyword.
Here, the keyword score determination means 23 determines the keyword score based on all of the translation scores related to a certain translation keyword, thereby increasing the keyword score of translation keywords corresponding to a plurality of input keywords at the same time. Can do. For example, the translation keyword “master” in FIG. 5 corresponds to both the input keywords “teacher” and “teacher”, and has a translation score other than 0 corresponding to each. However, the translation keyword “instructor” corresponds to the input keyword “teacher” but does not correspond to “teacher”, and the translation score for “teacher” is 0. As a result, the keyword score of the translation keyword “master” becomes higher. In this way, a more accurate search result can be obtained by setting a higher score for a more reliable translation keyword and placing importance on it, and setting a relatively lower score for the translation keyword.

Next, the document search means 24 uses the document search system 100 to search for a document based on the input keyword and the translation keyword, and obtains a plurality of documents as search result documents (step S4, document search step). In step S4, the document search means 24 passes the input keyword and translation keyword to the document database 120. The document database 120 includes all stored documents including any of the input keyword and translation keyword. The extracted document is returned to the document search means 24 as a search result document.
Here, the document search means 24 performs a search using an input keyword in a native language and a translation keyword in another language, so even a search from the document database 120 including documents in a plurality of languages produces a result in a single search. Obtainable.

  The search result document acquired in step S4 includes information (title, date, creator, etc.) for identifying the text data of the document, but may not necessarily include the text data. When the search result document does not include text data, the text data itself may be output from the document database 120 in response to a separate request from the user.

Each search result document is associated with information indicating the number of times each keyword appears in the text data.
FIG. 6 shows an example of such information. In this example, documents A to J are extracted as search result documents. For example, in the document A, the translation keyword “teacher” appears 12 times, the translation keyword “instructor” appears 10 times, and the translation keyword “master” appears 6 times. It is expressed that it becomes times. In this way, the document database 120 measures the number of times each input keyword and each translation keyword appears for each search result document, and returns it to the document search means 24 in association with each search result document. In FIG. 6, the search result documents are sorted in the order of the number of times each keyword appears.
Note that the document search device 10 may store the correspondence shown in FIG. 6 in the storage unit in the form of a table or the like.

In the example of FIG. 6, the number of times the keyword appears is used, but instead, the number of times the keyword appears may be added to the recognition rate by character recognition.
In a document in which a character string in a document is represented by a character code (text data, data for a word processor program, etc.), the number of occurrences of a keyword can be accurately counted using character code matching processing. On the other hand, in the case of a document in which a character string is represented by image data, it is necessary to perform character recognition processing to convert the image into a character code. However, the accuracy of this character recognition processing is not necessarily high. Therefore, at the time of character recognition processing, the degree of character recognition that can be performed on the document according to a predetermined standard may be evaluated as a recognition rate, and this recognition rate may be taken into account. For example, the value representing the number of appearances of the keyword may be decreased according to the recognition rate. Specifically, the number of occurrences of a keyword can be used as it is for a document with a recognition rate of 100%, and the number of appearances of a keyword can be halved for a document with a recognition rate of 50%.
Here, any method for calculating the recognition rate may be used as long as it is a known method in the conventional character recognition processing.

Next, the document score calculation means 25, for each search result document, the keyword score determined by the keyword score determination means 23 (see FIG. 5), and the number of times each of the input keyword and translation keyword appears (see FIG. 6). ) To calculate the document score (step S5, document score calculation step).
In this step S5, the document score is calculated, for example, by multiplying the keyword score of each keyword by the number of times that the keyword appears in the search result document, and adding up all the keywords. It can be said that this document score represents the possibility (probability) that the search result document is desired by the user.

FIG. 7 shows an example of a calculation result obtained by this calculation method. In the document A, the translated keyword “teacher” having the keyword score 90 appears 12 times, and the multiplication result is 90 × 12 = 1080. Similarly, the result of multiplication for the translation keyword “instructor” is 400, and the result of multiplication for the translation keyword “master” is 450. Other input keywords and translation keywords do not appear in the document A, and the result of multiplication is zero. The document score of the document A is a sum of these values, that is, 1930.
Note that the document search device 10 may store the correspondence shown in FIG. 7 in the storage unit in the form of a table or the like.
For a document in which a character string is represented by image data, the document score calculation means 25 calculates a document score based on the recognition rate in the character recognition process for the search result document in addition to the keyword score and the number of appearances. Also good.

Here, since the keyword score is different for each keyword, the document score of a document with many occurrences of the keyword does not necessarily increase. For example, the search result document having the highest number of occurrences of the keyword is the document A (28 times, see FIG. 6), but the document having the highest document score is the document C (2500, see FIG. 7). The order of has changed. This is because the keywords appearing in the document C are all input keywords, so the keyword score for each keyword is relatively high. Conversely, all the keywords appearing in the document A are translation keywords, so the keyword score for each keyword is relatively high. This is because it becomes lower. In addition, since the keyword scores are different among the translation keywords, more probable translation keywords are emphasized.
In this way, the document score calculation means 25 considers the quality of each keyword when calculating the document score of each search result document, so that it is compared with a method in which the document score is simply calculated based on the number of occurrences of the keyword. More accurate evaluation can be performed.

  Next, the search result output means 26 associates and outputs the search result documents (namely, documents A to J) and the document scores calculated by the document score calculation means 25 for each of them (step S6, search result output step). ). This output is made to the user via the display device 40, so that the user can know the search result. At this time, the search result output means 26 ranks the search result documents in descending order of the document score, and performs output according to this rank.

  As described above, the document search device 10, the document search method executed by the document search device 10, and the document search system 100 according to the first embodiment of the present invention are each configured to input keywords and translated keywords. Since the keyword score is determined and the document score is calculated based on the keyword score, the priority order of the documents output as the search result can be appropriately determined.

In the first embodiment described above, the language representing the input keyword is Japanese, and the languages representing the translation keyword are English and French, but these may be other languages, for example, Chinese. . The language representing the input keyword may be set according to the language used by the user, and the other language representing the translation keyword may be set according to the language of the document included in the document database 120.
The language representing the translation keyword may be a single language (for example, only English). Further, the translation service device 110 may output a single translation keyword with respect to an input keyword, or may output a plurality of translation keywords without ranking them. Even with such a configuration, if the keyword score is different between the input keyword and the translation keyword, a more accurate result than the conventional search can be obtained.

In the example of the first embodiment, OR search (logical sum search) is performed, and all documents in which any one of a plurality of input keywords and a plurality of translation keywords appear are acquired as search result documents. Unlike this, an AND search (logical product search) can also be executed.
In this case, in step S4 of FIG. 2, the document search means 24 passes the input keyword and the translation keyword to the document database 120 and instructs to perform an AND search. The document database 120 extracts all documents satisfying the following conditions i and ii from the stored documents, and returns the extracted documents to the document search unit 24 as search result documents.
-Condition i: For the input keyword "teacher", at least one of the input keyword itself and the corresponding translation keywords "teacher", "instructor", "master", "professeur", "instructeur" appears ii: For the input keyword “teacher”, at least one of the input keyword itself and the corresponding translation keywords “teacher”, “master”, “professeur” appears. In other words, the document search means 24 and the document database 120 For each input keyword, a keyword group for each input keyword is created by concatenating the input keyword and the corresponding translation keyword by an OR condition, and all the keyword groups are concatenated by an AND condition. Create search criteria

As a result of the search under this condition, for example, among the documents shown in FIG. 6 as the search result document in the first embodiment, the document H is the input keyword “teacher” and the corresponding translation keywords “teacher” and “master”. , “Professeur” is not included, so the condition ii is not satisfied and extraction is not performed. Similarly, the document J does not satisfy the condition ii and is not extracted.
In this example, the translation keywords “teacher”, “master”, and “professeur” are translation keywords corresponding to both of the two input keywords “teacher” and “teacher”. A document in which either appears will be extracted. For example, the document E includes the translation keyword “teacher”, and since this translation keyword satisfies both the condition i and the condition ii, the document E is extracted.
Even in the case of such an AND search, the processing after step S5 can be performed similarly to the OR search. That is, the document score is calculated and the search result is output as in the first embodiment. However, in this example, since the document H and the document J are not extracted in step S4, the processes after step S5 are not executed for the document H and the document J.

  In the first embodiment, the search using the translation keyword is always performed at the time of the search by the document search means 24. However, this may be switchable. For example, the search is performed using only the input keyword without using the translation keyword. It may be possible for the user to designate as appropriate. In this way, if necessary, it is possible to perform a process equivalent to a conventional document search using only input keywords.

The document database 120 may store, for each document to be searched, language information indicating in which language the document is expressed in association with each other, and the translation service device 110 similarly stores each of the translation keywords. May be stored in association with language information indicating in which language the translated keyword is expressed. In this case, the input keyword may always be handled in a predetermined language corresponding to the mother tongue.
For example, even if a Japanese keyword is translated into Chinese, the same notation (a character string represented by the same character code) may be obtained. For such keywords, the keyword score of the input keyword can be applied to a document in Japanese, and the keyword score of a translation keyword can be applied to a document in Chinese. That is, for keywords that have the same notation in different languages among input keywords and translation keywords, the keyword score of the search result document that matches the language information is used when calculating the document score of the search result document. Also good.
In this way, even if keywords having the same notation in a plurality of languages are included, the likelihood of each keyword can be appropriately evaluated.

  In the first embodiment, the document database 120 measures the number of times a keyword appears in a search result document, but this may be measured by other components. For example, the text data of the search result document may be transferred from the document database 120 to the document search device 10 and measured by the document search unit 24 or the document score calculation unit 25 of the document search device 10.

  The translation service device 110 and the document database 120 may be anything as long as appropriate information can be exchanged with the document search device 10 regarding keyword translation and document search. The functions as the translation service device 110 and the document database 120 may be realized by executing a program stored in each storage unit. In this case, the program of the document search apparatus 10, the program of the translation service apparatus 110, and the program of the document database 120 cause these computers to function as the document search system 100 as a document search program.

  In the hardware configuration of the first embodiment, the document search apparatus 10 that is a single computer includes a keyword receiving unit 21, a keyword translating unit 22, a keyword score determining unit 23, a document searching unit 24, a document score calculating unit 25, and a search. The result output means 26 is included, and separately from this, the translation service device 110 and the document database 120 are provided as a single computer. However, the hardware configuration may be different. For example, a computer constituting the document search apparatus 10 may have a function as the translation service apparatus 110 and a function as the document database 120.

Claims (8)

A document search device for searching for documents using keywords;
A translation service device;
A document search system including a document database for storing a plurality of documents to be searched,
The document search device includes:
Keyword accepting means for receiving one or more keywords as input keywords;
Corresponding to each of the input keywords, keyword translation means for acquiring a translation keyword obtained by translating the input keyword into another language for a plurality of other languages;
Keyword score determining means for determining a keyword score for each of the input keyword and the translated keyword;
Document search means for searching for a document based on the input keyword and the translation keyword and obtaining a plurality of search result documents;
Document score calculation means for calculating a document score based on the keyword score for each of the search result documents;
Search result output means for associating and outputting each of the search result documents and the corresponding document score,
The translation service device generates a plurality of translation keywords having a rank corresponding to each of the input keywords,
The translation service device represents the ranking of the translation keywords based on the posting order of the translation keywords,
The keyword score determining means determines a translation score based on the rank for all combinations of each of the input keywords and each of the translation keywords,
The keyword score determining means determines the keyword score based on all of the related translation scores for each of the translation keywords,
The keyword score of the input keyword is higher than any of the keyword scores of the translation keyword corresponding to the input keyword.
Document search system. The document database stores, in association with each of the documents, language information indicating in which language the document is expressed,
The translation service device associates and stores language information indicating in which language the translation keyword is expressed for each of the translation keywords,
The translation service device stores language information representing a predetermined language corresponding to the input keyword,
The document score calculation means applies the keyword score of the input keyword and the translation keyword when the language information matches the language information of the document when calculating the document score of the document. Item 2. The document search system according to Item 1. The document search system accepts designation as to whether or not to use the translation keyword,
In response to the designation, the document score calculation means
The document score is calculated based on the input keyword and the translation keyword, or the document score is calculated based on only the input keyword regardless of the translation keyword. Document search system.   The document score calculation means further calculates the document score based on the number of times each of the input keyword and the translation keyword appears in the search result document. Document search system.   The document search system according to claim 4, wherein the document score calculation means further calculates the document score based on a recognition rate in a character recognition process for the search result document. The document search means comprises a function of searching for a document by OR search and a function of searching for a document by AND search when the keyword receiving means receives two or more input keywords,
The document search means searches for a document in which either the input keyword or the translation keyword appears in the OR search,
The document search means searches for documents satisfying a condition that “at least one of the input keyword and each translation keyword corresponding to the input keyword appears” for all the input keywords in the AND search. The document search system according to any one of 1 to 5.   A document search program for causing a computer to function as the document search system according to any one of claims 1 to 6.   A document search method for searching for a document by the document search system according to claim 1.

Images