Device and method for retrieving information and recording medium

Description: TECHNICAL FIELD [Detailed description of the invention]

【0001】

[Technical field to which the invention belongs]

The present invention relates to an information retrieval device, an information retrieval method, and a recording medium that perform information retrieval based on a designated word or compound word.

【0002】

[Conventional technology]

With the miniaturization of computers and the spread of the Internet, the situation of creating and browsing documents on a computer is increasing, and software for elaborating documents and looking up dictionaries on a computer has already been put into practical use. ing.

【0003】

For example, in dictionary search software, dictionary data is stored on a CD-ROM or the like, and when the user is creating a document or reading a document, the character string for which detailed information is desired is input as the search character string. This makes it possible to refer to the information stored in the CD-ROM or the like.

【0004】

[Problems to be Solved by the Invention]

However, in the conventional data search software, only a method of searching and extracting the data stored in the CD-ROM or the like and presenting all the successfully extracted data to the user has been adopted. In this case, in order to obtain the desired information, the user had to manually search for the information he / she needed from the data, which was a heavy burden on the user.

【0005】

The present invention has been made in consideration of the above circumstances, and provides an information retrieval device, an information retrieval method, and a recording medium that reduce the burden of the user narrowing down the necessary information from the search results that have been required so far. The purpose is.

【0006】

[Means for solving problems]

The information retrieval device according to the present invention (claim 1) is based on a means for designating character data of a word or a compound word in a document and at least a portion having a predetermined relationship with the designated character data in the document. , A means for obtaining one or more keywords, and a means for determining data to be presented as a search result for a predetermined database based on at least the specified character data and the obtained keywords. And.

【0007】

The information retrieval device according to the present invention (claim 2) obtains target data from a predetermined database based on a means for designating character data of a word or a compound word in a document and the character data designated by this means. Specific data from the search results searched from the database based on the means for searching, the means for extracting at least one sentence including the character data and analyzing the sentence, and the analysis result of the sentence by this means. It is characterized by having a means for selecting.

【0008】

Preferably, the analysis result of the sentence may be a keyword extracted by the morphological analysis result of the sentence.

【0009】

Preferably, the analysis result of the sentence may be selected from the keywords extracted by the morphological analysis result of the sentence, which are close in distance.

【0010】

The information retrieval device according to the present invention (claim 5) cuts out a means for designating character data of a word or a compound word in a document and at least one sentence including the character data designated by this means, and analyzes the sentence. A means for extracting keywords by row, a means for creating a search condition expression from the keywords extracted by this means, and a means for searching target data from a predetermined database based on the search conditions created by this means. It is characterized by having and.

【0011】

Preferably, when creating a search condition from the extracted keywords, the search condition may be created using only the extracted keywords that are close in distance.

【0012】

Preferably, other related keywords may be extracted from the extracted keywords using a predetermined database, and at least one of those keywords may be used to select specific data.

【0013】

The information retrieval method according to the present invention (claim 8) specifies character data of a word or a compound word in a document, and is based on at least a portion having a predetermined relationship with the specified character data in the document. Alternatively, a plurality of keywords are obtained, and at least based on the specified character data and the obtained keywords, data to be presented as a search result for a predetermined database is determined.

【0014】

In the information retrieval method according to the present invention (claim 9), character data of a word or a compound word in a document is specified, target data is searched from a predetermined database based on the specified character data, and the character is described. It is characterized in that at least one sentence including data is cut out, the sentence is analyzed, and specific data is selected from the search results searched from the database based on the analysis result of the analyzed sentence.

【0015】

In the information retrieval method according to the present invention (claim 10), character data of a word or a compound word in a document is specified, at least one sentence including the specified character data is cut out, and the sentence is analyzed to extract a keyword. Then, a search condition expression is created from the extracted keywords, and the target data is searched from a predetermined database based on the created search conditions.

【0016】

According to the present invention (claim 11), a computer is made to input character data of a word or a compound word in a document, and at least based on a portion having a predetermined relationship with the specified character data in the document, 1 Or a computer-readable record of a program that asks for multiple keywords and, at least based on the specified character data and the asked keywords, determines the data to be presented as search results for a given database. The gist is the medium.

【0017】

The present invention (claim 12) receives input of character data of a word or a compound word specified in a document, searches for target data from a predetermined database based on the character data, and includes the character data. An information search characterized by having a function of cutting out at least one sentence, analyzing the sentence, and selecting specific data from the search results searched from the database based on the analysis result of the sentence. The gist is the recording medium for.

【0018】

The present invention (claim 13) receives input of character data of a word or a compound word specified in a document, cuts out at least one sentence including the character data, analyzes the sentence, extracts a keyword, and extracts the keyword. The gist is a recording medium for information retrieval, which is characterized by having a function of creating a search condition expression from the created keywords and searching the target data from a predetermined database based on the created search conditions. To do.

【0019】

The present invention relating to the device is also valid as an invention relating to the method, and the present invention relating to the method is also valid as an invention relating to the device.

【0020】

In addition, the present invention relating to an apparatus or method provides a function for causing a computer to perform a procedure corresponding to the present invention (or for causing the computer to function as a means corresponding to the present invention, or for causing the computer to perform a function corresponding to the present invention. It can also be used as a computer-readable recording medium on which a program (to be realized) is recorded.

【0021】

In the present invention, information before and after the character string specified by the user can be automatically extracted, and the extracted information can be used to narrow down the information to be displayed when displaying the search result.

【0022】

For example, if you specify character data such as words and compound words for which you want to know related or detailed information in a document, the target data is searched from a predetermined database based on the specified character data, and the character data is searched for. At least one sentence including the sentence is cut out, the sentence is analyzed, and specific data is selected from the search results searched from the database based on the analysis result of the analyzed sentence. Alternatively, for example, at least one sentence containing the specified character data is cut out, the sentence is analyzed to extract keywords, a search condition expression is created from the extracted keywords, and based on the created search conditions, the search condition expression is created. Search the target data from the specified database.

【0023】

According to the present invention, the user can efficiently obtain necessary information, which is effective in accelerating the understanding of sentences. In other words, since the information is presented without the user narrowing down the necessary information by himself / herself, the load on the user is reduced.

【0024】

BEST MODE FOR CARRYING OUT THE INVENTION

Hereinafter, embodiments of the invention will be described with reference to the drawings.

【0025】

(First Embodiment) The first embodiment of the present invention will be described.

【0026】

FIG. 1 shows a configuration example of the information retrieval device according to the present embodiment. As shown in FIG. 1, the information retrieval device includes an input unit 101, a search control unit 102, a sentence analysis unit 103, a database search unit 104, an analysis knowledge 105, a database 106, and an output unit 107.

【0027】

The input unit 101 is for inputting a document, and also for inputting an instruction of a word or idiom to be searched in the input document, an instruction of selecting the searched display information, and the like. ..

【0028】

Document input includes, for example, keyboard (key input), recording medium drive device (reading from a recording medium such as a magnetic disk, magnetic tape, or optical disk), OCR (optical reading), network connection device (acquisition by communication), etc. Use one or more of the above. When using OCR, the function of recognizing the read character image and generating a code string is used.

【0029】

For example, a keyboard, a mouse, or the like is used for inputting an instruction from the user.

【0030】

The document data input from the input unit 101 is temporarily stored in the memory (not shown) of the search control unit 102, and then the search for certain character data (words and compound words) in the document being displayed via the input unit 101. Is instructed, a predetermined part of the document data (for example, one sentence including the instructed character data or a plurality of sentences including at least the one sentence) is passed to the sentence analysis unit 103.

【0031】

The character data instructed to be searched by the input unit 101 is input to the database search unit 104 via the search control unit 102.

【0032】

Under the control of the search control unit 102, the sentence analysis unit 103 receives character data from the search control unit 102, performs sentence analysis using the analysis knowledge 105, and outputs an analysis result. As a method of sentence analysis, general methods such as morphological analysis, syntactic analysis, and semantic analysis can be used.

【0033】

Under the control of the search control unit 102, the database search unit 104 receives character data from the search control unit 102, refers to the database 106, and outputs a search result.

【0034】

The search control unit 102 selects the data requested by the user by using the sentence analysis result by the sentence analysis unit 103 and the search result of the database 106 by the database search unit 104, and passes it to the output unit 107.

【0035】

The output unit 107 is for presenting the input document and the search result to the user.

【0036】

As the output unit 107, any one or a plurality of types such as an optical display and a printer are used.

【0037】

Hereinafter, the present embodiment will be described in more detail with reference to specific examples.

【0038】

Here, it is assumed that the document has already been input and displayed. The user instructs the character string (hereinafter, referred to as a search character string) for which detailed information in the document is to be obtained by the input unit 101.

【0039】

As a specific example, as illustrated in FIG. 2, the user inputs a mouse or the like on a display screen on which a document such as ".... He is the director of this movie ..... ...." is displayed. Suppose you use Part 101 to drag the string part "Director" and instruct it to get more information about "Director".

【0040】

The search control unit 102 sends one sentence including the search character string instructed by the user to the cut-out sentence analysis unit 103 using the position information or the like. Also, the search character string is sent to the database search unit 104.

【0041】

In the example of FIG. 2, one sentence including the search character string "director", "he is the director of this movie" is cut out and sent to the sentence analysis unit 103. Further, the search character string "supervisor" is sent to the database search unit 104.

【0042】

The sentence analysis unit 103 performs analysis processing on one sentence input from the search control unit 102 using the analysis knowledge 105, and passes the analysis result to the search control unit 102. The analysis process referred to here refers to all or a part of morphological analysis, syntactic analysis, semantic analysis, and the like. Here, as an example, a case where morphological analysis is used will be described. FIG. 3 shows an example of the analysis result when the morphological analysis is performed on the above sentence, that is, "He is the director of this movie."

【0043】

The database search unit 104 extracts the target data from the database 106 using the search character string "supervisor" input from the search control unit 102, and passes it to the search control unit 102.

【0044】

For example, if the database 106 is like the Japanese dictionary illustrated in FIG. 4, the search character string "director" passed from the search control unit 102 is used to use the word "on a person". Standing instructions, people, institutions, "people who direct in movies, plays, etc." are taken out.

【0045】

Here, the part of the data in FIG. 4 surrounded by <> is a tag. For example, the part narrowed from "<heading word>" to "<\ heading word>" becomes the heading word of the dictionary. In this example, the dictionary entry word is "director". Next, the part of speech information for the headword is described. Here, it is shown that it is a noun used as a verb that is associated with the noun and the verb "suru" to utilize the s-irregular conjugation. Next, the meaning of the headword is described. Here, there are three meanings, each of which is described in "word meaning part 1" to "word meaning part 3".

【0046】

In the present embodiment, the processing by the sentence analysis unit 103 and the processing by the database search unit 104 may be executed at the same time, or one of them may be executed first and the other may be executed later.

【0047】

The analysis result from the sentence analysis unit 103 and the search result from the database search unit 104 are returned to the search control unit 102, and the content to be displayed to the user is selected.

【0048】

Here, FIG. 5 shows a configuration example of the search control unit 102 in the present embodiment. As shown in FIG. 5, the search control unit 102 includes a control storage unit 601, a keyword extraction unit 602, and a display content selection unit 603.

【0049】

The control storage unit 601 is involved in the entire search, such as storing a document input from the input unit 101, receiving a search instruction from a user and sending it to the database search unit 104, or cutting out one sentence and sending it to the sentence analysis unit 103. Take control.

【0050】

The keyword extraction unit 602 extracts keywords effective for selecting display contents from the analysis results passed from the sentence analysis unit 103, creates a keyword list, and passes it to the control storage unit 601. The control storage unit 601 passes the keyword list and the search results (see FIG. 4) from the database search unit 104 to the display content selection unit 603.

【0051】

FIG. 6 shows an example of the procedure of the keyword list creation process of the keyword extraction unit 602.

【0052】

First, set the variable n to 1 (step S702). This value represents the distance from the search string (in the case of Figure 3, "supervisor"). The distance mentioned here is a value of how many words are separated from the search character string.

【0053】

The morphological analysis result is determined whether or not it becomes a word-for-word keyword (steps S703, S704). The keywords referred to here are for narrowing down the search results, and refer to terms extracted by using part-speech information such as whether it is a noun or a proper noun, or field information such as computer terms and economic terms as keys. Here, the key for keyword extraction is used as a noun. The keyword words are stored in the keyword list (step S705).

【0054】

In addition, determine if the variable n is out of scope (step S706). The scope referred to here is a value indicating how many words away from the search character string are targeted for keyword extraction. For example, if the scope is 5, keywords will be extracted from the 5 words before and after the search string. In this embodiment, the scope is 3. If the scope is not exceeded, increase n by 1 (step S707) and search for the keyword again. If it exceeds, all the words in the scope have been searched, so the process is terminated (step S708), and the keyword list as the process result is passed to the display content selection unit 603. If the keyword is not extracted, an empty keyword list is passed to the display content selection unit 603.

【0055】

By the above processing, for example, when the morphological analysis result shown in FIG. 3 is input, "movie" is extracted as a keyword and passed to the display content selection unit 603.

【0056】

The display content selection unit 603 receives the keyword list passed from the keyword extraction unit 602 and the search result passed from the database search unit 104, and calculates an evaluation point for the search result in order to select the display content. , The one with the highest score is passed to the output unit 107.

【0057】

Figure 7 shows an example of the procedure as a display content selection method. In this example, the number of occurrences of a keyword in each word meaning is checked, and the one with the highest number of occurrences is displayed to the user. In this specific example, it is assumed that the display content selection unit 603 receives the search result represented by FIG. 4 and the keyword list movie from the control storage unit 601.

【0058】

First, the keyword number table is initialized (step S902). The keyword number table is a table that stores the number of keywords appearing in the meaning data, and prepares as many as the number of meaning data and makes a one-to-one correspondence. When the search result shown in FIG. 4 is received, since this search result contains three meanings, three keyword number tables are prepared as shown in FIG.

【0059】

Then set the variable n to 1 (step S903). This value represents the entry number of the semantic data.

【0060】

The word meaning data is searched for a character string in the keyword list (steps S904, S905). If there are keywords, the number of keywords is counted and assigned to the keyword number table that has a one-to-one correspondence with the semantic data (step S906).

【0061】

It is determined whether or not there is the next meaning data (step S907), and if there is, n is incremented by 1 (step S908), and the keyword search in the meaning data is performed again. If there is no keyword, the process ends (step S909), the word meaning with the most keywords is sent to the output unit 107 using the keyword number table, and the dictionary search result is displayed.

【0062】

In this specific example, since the keyword "movie" does not exist in the first word meaning data, the first keyword number table is 0 as shown in FIG. Since there are two words "movie" in the second word meaning data, 2 is assigned to the second keyword number table (step S906). In the same manner below, the number of keywords is calculated for each word meaning data. As a result, the second word meaning data in FIG. 4 is selected as the most appropriate word meaning.

【0063】

Then, for example, as shown in FIG. 10, the output unit 107 displays the dictionary search result.

【0064】

In the display example of FIG. 10, only the data most related to the search character string is displayed, but various other variations can be considered in the display form of the dictionary search result.

【0065】

For example, other data (for example, the first or third semantic data in FIG. 4) may be displayed by clicking the pop-up display screen with a mouse or the like. The next candidate can be selected by prioritizing each candidate by using or combining methods such as the one with the next highest number of occurrences of keywords and the one with the smaller entry number.

【0066】

FIG. 11 shows a display example when the user instructs the display of the next candidate in the above specific example. In the example of FIG. 11, the priority order of giving priority to the one with the smaller entry number is used.

【0067】

Further, for example, as shown in FIG. 12, instead of selecting and displaying one search result, all the search results may be sorted and displayed in order of priority.

【0068】

In the present embodiment, a scope based on the number of words that are far from the search target word is set as the keyword search range, and the words are within the scope of the search range. You may use a method such as cutting out one sentence in one sentence and searching the sentences before and after the sentence if it is not found in the sentence.

【0069】

Further, various databases 106 can be used. For example, if the database 106 is like a synonym dictionary, synonyms related to "supervision" can be selected and displayed as illustrated in FIG. it can.

【0070】

As described above, according to the present embodiment, the search results are not presented as they are, but only those that are expected by the user are selected and presented, or even when all the search results are presented. It is possible to take a presentation method such as presenting in a form that informs which one is expected to be more desired by the user. As a result, the user can efficiently obtain the necessary information. That is, since the information is presented without narrowing down the necessary information by the user, the understanding of the sentence is accelerated, and the load on the user can be reduced as compared with the conventional case.

【0071】

(Second Embodiment) Next, the second embodiment of the present invention will be described.

【0072】

In the first embodiment, the case where the morphological analysis is used as the analysis result of the sentence analysis unit 103 has been described, but in the present embodiment, the case where the syntax analysis is used as the analysis result of the sentence analysis unit 103 will be described.

【0073】

In the following, the differences from the first embodiment will be mainly described.

【0074】

First, the process according to the input of the search instruction by the user and the process by the database search unit 104 are the same as those in the first embodiment.

【0075】

Here, FIG. 14 shows an example of the output result when the morphological analysis / syntax analysis is used as the analysis result of the sentence analysis unit 103. When the analysis result of the statement analysis unit 103 is the syntax analysis result (concept-dependent structure data) as shown in FIG. 14, the information when the syntax analysis result is used by changing the keyword extraction unit 602 of the search control unit 102. A search device can be realized.

【0076】

FIG. 15 shows an example of the processing procedure of the keyword extraction unit 602 in the search control unit 102 when using the parsing result.

【0077】

First, set the variable n to 1 (step S802). This value represents the distance from the node that indicates the search string. The distance mentioned here is a value of how many times the link is followed from the search character string. That is, when viewed from the "director" node in FIG. 14, the "movie" and "is" nodes have a distance of 1, and the "his" node has a distance of 2.

【0078】

The parsing result is determined to be a word-for-word keyword (steps S803, S804). The keywords referred to here are for narrowing down the search results, and refer to terms extracted by using part-speech information such as whether it is a noun or a proper noun, or field information such as computer terms and economic terms as keys. Here, the key for keyword extraction is used as a noun. The keyword words are stored in the keyword list (step S805).

【0079】

In addition, determine if the variable n is out of scope (step S806). The scope referred to here is a value indicating how far the distance from the node indicating the search character string is to be the target of keyword extraction. For example, if the scope is 5, keywords will be extracted from the nodes whose distance from the node indicating the search character string is 5 or less. In this embodiment, the scope is 2. If the scope is not exceeded, increase n by 1 (step S807) and search for the keyword again. If it exceeds, all the words in the scope are searched, so the processing is terminated, and the keyword list which is the processing result is passed to the display content selection unit 603. If the keyword is not extracted, an empty keyword list is passed to the display content selection unit 603.

【0080】

By the above processing, for example, when the syntax analysis result shown in FIG. 14 is input, "movie" is extracted as a keyword and passed to the display content selection unit 603.

【0081】

After that, as in the first embodiment, the display content selection unit 603 selects the data to be presented based on the keyword list passed from the keyword extraction unit 602 and the search result passed from the database search unit 104. For example, as shown in FIG. 10, the output unit 107 displays the dictionary search result.

【0082】

The present invention is not limited to the first and second embodiments described above. For example, as sentence analysis, we have given examples using morphological analysis and syntactic analysis, but you can simply search for keywords. Further, as the database, the Japanese dictionary is used in the above, but other foreign word dictionaries, new word dictionaries, English-Japanese dictionaries and the like may be used.

【0083】

(Third Embodiment) Next, the third embodiment of the present invention will be described.

【0084】

In the first embodiment, the search results obtained by searching the database 106 based on the designated words and idioms are further narrowed down by the keywords extracted from the sentence analysis results. In, the search results for the database 106 are obtained based on the keywords (main search conditions) extracted from the sentence analysis results and the specified words and idioms (subordinate search conditions).

【0085】

The configuration example of the information retrieval device according to the present embodiment is the same as that of the first embodiment. That is, as shown in FIG. 1, the information retrieval device includes an input unit 101, a search control unit 102, a sentence analysis unit 103, a database search unit 104, an analysis knowledge 105, a database 106, and an output unit 107.

【0086】

In the present embodiment, the processing content in the search control unit 102 is different from that in the first embodiment, but other than that, it is basically the same as in the first embodiment.

【0087】

In the following, the differences from the first embodiment will be mainly described.

【0088】

The document data input from the input unit 101 is temporarily stored in the memory (not shown) of the search control unit 102, and then the search for certain character data (words and compound words) in the document being displayed via the input unit 101. Is instructed, a predetermined part of the document data (for example, one sentence including the instructed character data or a plurality of sentences including at least the one sentence) is passed to the sentence analysis unit 103.

【0089】

Under the control of the search control unit 102, the sentence analysis unit 103 receives character data from the search control unit 102, performs sentence analysis using the analysis knowledge 105, and outputs an analysis result. As a method of sentence analysis, general methods such as morphological analysis, syntactic analysis, and semantic analysis can be used.

【0090】

The search control unit 102 creates a search condition expression for obtaining the data requested by the user by using the character data instructed to search by the input unit 101 and the sentence analysis result passed from the sentence analysis unit 103. Then, it is given to the database search unit 104 to instruct the search.

【0091】

Under the control of the search control unit 102, the database search unit 104 receives the search conditional expression passed from the search control unit 102, refers to the database 106, and outputs a search result.

【0092】

The search control unit 102 passes the search result of the database 106 by the database search unit 104 to the output unit 107.

【0093】

Hereinafter, the present embodiment will be described in more detail with reference to specific examples.

【0094】

Here, it is assumed that the document has already been entered. The user uses the input unit 101 to specify a character string (that is, a search character string) for which detailed information is to be obtained in the document.

【0095】

As a specific example, as illustrated in FIG. 16, the user inputs a mouse or the like on a display screen on which a document such as ".... He is the director of Jurassic World ..... ...." is displayed. Suppose you use the device to drag the string part of "supervisor" and instruct you to get detailed information about "supervisor". In addition, "Jurassic World" shall be a proper noun representing the movie name.

【0096】

The search control unit 102 sends one sentence including the search character string instructed by the user to the cut-out sentence analysis unit 103 using the position information or the like.

【0097】

In the example of FIG. 16, one sentence including the search character string director, he is the director of Jurassic World is cut out and sent to the sentence analysis unit 103.

【0098】

The sentence analysis unit 103 performs analysis processing on one sentence passed from the search control unit 102 by using the analysis knowledge 105, and passes the analysis result to the search control unit 102. The analysis process referred to here refers to all or part of morphological analysis, syntactic analysis, semantic analysis, and the like. Here, as in the first embodiment, a case where morphological analysis is used will be described as an example. FIG. 17 shows an example of the analysis result when the morphological analysis is performed on the above sentence, that is, "He is the director of Jurassic World."

【0099】

The search control unit 102 creates a search conditional expression using the analysis result from the sentence analysis unit 103 and the search character string, and passes it to the database search unit 104.

【0100】

Here, FIG. 18 shows a configuration example of the search control unit 102 in the present embodiment. As shown in FIG. 18, the search control unit 102 includes a control storage unit 2501 and a search condition expression creation unit 2502.

【0101】

The control storage unit 2501 stores the document input from the input unit 101, receives a search instruction from the user, cuts out one sentence and sends it to the sentence analysis unit 103, and sends the database search result to the output unit 107. It controls the entire search.

【0102】

The search condition expression creation unit 2502 extracts keywords effective for searching a dictionary from the analysis results from the sentence analysis unit 103, creates a search condition expression, and passes it to the database search unit 104.

【0103】

FIG. 19 shows an example of the procedure for creating a search conditional expression by the search conditional expression creation unit 2502.

【0104】

First, the variable n is set to 1 (step S2602), and words (distances) n words away from the search string (director in the case of FIG. 17) are extracted (step S2603).

【0105】

The morphological analysis result is determined whether or not it becomes a word-for-word keyword (step S2604). The keywords mentioned here are effective for creating search condition expressions, and are extracted using part of speech information such as nouns or proper nouns and field information such as computer terms and economic terms as keys. Refers to a term. Here, the key for keyword extraction is a proper noun. The keyword words are stored in the keyword list (step S2605).

【0106】

In addition, determine if the variable n is out of scope (step S2606). The scope referred to here is a value indicating how many words away from the search character string are targeted for keyword extraction. For example, if the scope is 5, keywords will be extracted from the 5 words before and after the search string. In this embodiment, the scope is 3. If the scope is not exceeded, increase n by 1 (step S2607) and search for the keyword again. If it exceeds, a search condition expression is created from the keyword list and the search character string (step S2608), the process is terminated (step S2609), and the search condition expression is passed to the database search unit 104. As a method for creating a search condition expression from a keyword list and a search character string, a known method may be used.

【0107】

By the above processing, for example, when the morphological analysis result shown in FIG. 17 is input, "Jurassic World" is extracted as a keyword, and a search for searching "Jurassic World Director" together with the search character string "Director". The conditional expression is created as shown in FIG.

【0108】

The database search unit 104 extracts the target data from the database using the search conditional expression passed by the search control unit 102 and passes it to the search control unit 102.

【0109】

For example, assuming that the database 106 is like the movie encyclopedia illustrated in FIG. 21, information about "Jurassic World" is extracted from the search conditional expression shown in FIG. 20 passed from the search control unit 102, and the information is extracted from the conditional expression. Since the director is required as a feature, the character string "Steve " is taken out.

【0110】

Here, the part of the data in FIG. 21 surrounded by <> is a tag. For example, the part narrowed from "<heading word>" to "<\ heading word>" is the heading word of the encyclopedia. In this example, the headword of the encyclopedia is "Jurassic World". Next, the year of publication for the headword is described. Here, it is shown that it was released in 1993. Next, it is described as the creator and director of the headword. The description of the description will be omitted below. In addition, "Steve " etc. shall represent a person's name.

【0111】

The searched character string is passed to the output unit 107 via the search control unit 102, and the dictionary search result is displayed.

【0112】

For example, as shown in FIG. 22, the dictionary search result is presented to the user by the output device of the output unit 107.

【0113】

In the above, a search condition expression is created from the keyword list and the search character string, and the search results obtained by searching the database using this search condition expression are presented. Instead, the search is performed from the keyword list. A conditional expression may be created, the database may be searched by the search conditional expression, and then the search result corresponding to the search character string may be selected and presented.

【0114】

(Fourth Embodiment) Next, the fourth embodiment of the present invention will be described.

【0115】

In the third embodiment, the case where the morphological analysis is used as the analysis result of the sentence analysis unit 103 has been described, but in the present embodiment, the case where the syntax analysis is used as the analysis result of the sentence analysis unit 103 will be described.

【0116】

In the following, the differences from the third embodiment will be mainly described.

【0117】

First, when a search instruction is input by the user, for example, one sentence including the search character string "director", that is, "he is the director of Jurassic World" is cut out and sentence analysis is performed as in the first embodiment. Sent to department 103.

【0118】

Here, FIG. 23 shows an example of the output result when the morphological analysis / syntax analysis is used as the analysis result of the sentence analysis unit 103. When the analysis result of the statement analysis unit 103 is the syntax analysis result (concept-dependent structure data) as shown in FIG. 23, when the syntax analysis result is used by changing the search condition expression creation unit 2502 of the search control unit 102. Search device can be realized.

【0119】

FIG. 24 shows an example of the processing procedure of the search condition expression creation unit 2502 in the search control unit 102 when the syntactic analysis result is used.

【0120】

First, set the variable n to 1 (step S2702).

【0121】

Extract the value n of how many times the link was followed from the search string (step S2703). This value n represents the distance from the node indicating the search string. That is, when viewed from the "director" node in FIG. 23, the "Jurassic World" and "is" nodes have a distance of 1, and the "his" node has a distance of 2.

【0122】

The parsing result is determined to be a word-for-word keyword (step S2704). The keywords referred to here are for narrowing down the search results, and refer to terms extracted by using part-speech information such as whether it is a noun or a proper noun, or field information such as computer terms and economic terms as keys. Here, the key for keyword extraction is a proper noun. The keyword words are stored in the keyword list (step S2705).

【0123】

In addition, determine if the variable n is out of scope (step S2706). The scope referred to here is a value indicating how far the distance from the node indicating the search character string is to be the target of keyword extraction. For example, if the scope is 5, keywords will be extracted from the nodes whose distance from the node indicating the search character string is 5 or less. In this embodiment, the scope is 2. If the scope is not exceeded, increase n by 1 (step S2707) and search for the keyword again. If it exceeds, all the words in the scope have been searched, so a search condition expression is created from the keyword list and the search string (step S2708), the processing is finished (step S2709), and the search condition expression is in the database. It is passed to the search unit 104. As for the method of creating a search condition expression from a keyword list and a search character string, a method generally used in the current technology may be used.

【0124】

By the above processing, for example, when the parsing result shown in FIG. 23 is input, "Jurassic World" is extracted as a keyword, and a search for searching "Jurassic World Director" together with the search character string "Director" is performed. The conditional expression is created as shown in FIG.

【0125】

After that, as in the third embodiment, the database search unit 104 retrieves the target data from the database using the above search conditional expression. For example, as shown in FIG. 22, the output device of the output unit 107. Presents the dictionary search results to the user.

【0126】

The present invention is not limited to the third and fourth embodiments described above. For example, sentence analysis uses morphological analysis and syntactic analysis, but you can simply search for keywords. Further, as the database, the Japanese dictionary is used in the above, but other dictionaries such as a Japanese dictionary, a loanword dictionary, a new word dictionary, and an English-Japanese dictionary may be used.

【0127】

Further, in the first to fourth embodiments, one sentence including the search character string instructed by the user is cut out as the target of sentence analysis for keyword extraction, but how to cut out the sentence to be the target of this sentence analysis. In addition to this, various methods can be considered. For example, one sentence including the search character string instructed by the user and a plurality of sentences before and after the sentence may be cut out as a group.

【0128】

(Fifth Embodiment) Next, the fifth embodiment of the present invention will be described.

【0129】

In the first to fourth embodiments, the keyword obtained from the sentence analysis result is used as it is for the search or the selection of the search result, but in the present embodiment, another 1 or another one or the keyword obtained from the sentence analysis result is used. A plurality of keywords are requested, and the latter keyword is used instead of the former keyword or in combination with the former keyword.

【0130】

This configuration can be applied to any of the first to fourth embodiments, but in the following, in the first embodiment described above, other keywords extracted from the sentence analysis result are used using a predetermined database. A case where related keywords are extracted and specific data is selected from the search results by using at least one of the keywords will be described.

【0131】

In the following, the differences from the first embodiment will be mainly described.

【0132】

The configuration example of the information retrieval device according to the present embodiment is the same as that of the first embodiment. That is, as shown in FIG. 1, the information retrieval device includes an input unit 101, a search control unit 102, a sentence analysis unit 103, a database search unit 104, an analysis knowledge 105, a database 106, and an output unit 107.

【0133】

The present embodiment is different from the first embodiment in that the above-mentioned function is added to the search control unit 102, but other than that, it is basically the same as the first embodiment.

【0134】

Hereinafter, the present embodiment will be described in more detail with reference to specific examples.

【0135】

Here, it is assumed that the document has already been entered. The user uses the input unit 101 to specify a character string (that is, a search character string) for which detailed information in the document is to be obtained.

【0136】

As a specific example, as illustrated in FIG. 16, the user inputs a mouse or the like on a display screen on which a document such as ".... He is the director of Jurassic World ..... ...." is displayed. Suppose you use the device to drag the string part of "supervisor" and instruct you to get detailed information about "supervisor". In addition, "Jurassic World" shall be a proper noun representing the movie name.

【0137】

The search control unit 102 cuts out a sentence including the search character string instructed by the user using the position information and the like, and sends it to the sentence analysis device 103.

【0138】

In the example of FIG. 16, a sentence including the search character string director, he is the director of Jurassic World is cut out and sent to the sentence analysis unit 103.

【0139】

The sentence analysis unit 103 performs analysis processing on one sentence passed from the search control unit 102 using the analysis knowledge 105, and passes the analysis result to the search control unit 102. The analysis process referred to here refers to all or part of morphological analysis, syntactic analysis, semantic analysis, and the like. Here, as in the first embodiment, a case where morphological analysis is used will be described as an example. FIG. 17 shows an example of the analysis result when the morphological analysis is performed on the above sentence, that is, "He is the director of Jurassic World."

【0140】

The database search unit 104 extracts the target data from the database 106 using the search character string "supervisor" input from the search control unit 102 and passes it to the search control unit 102.

【0141】

Here, it is assumed that the database 106 is like a Japanese dictionary and the results as illustrated in FIG. 4 are extracted as in the first embodiment.

【0142】

The analysis result from the sentence analysis unit 103 and the search result from the database search unit 104 are returned to the search control unit 102, and the content to be displayed to the user is selected.

【0143】

Here, FIG. 25 shows a configuration example of the search control unit 102 in the present embodiment. As shown in FIG. 25, the search control unit 102 includes a control storage unit 2401, a keyword extraction unit 2402, a database 2403, and a display content selection unit 2404.

【0144】

The control storage unit 2401 is the same as that of the first embodiment, and stores the document input from the input unit 101, receives a search instruction from the user and sends it to the database search unit 104, cuts out one sentence and analyzes the sentence. Controls the entire search, such as sending to section 103.

【0145】

The keyword extraction unit 2402 extracts keywords effective for selecting the content to be presented from the analysis results passed from the sentence analysis unit 103 while referring to the database 2403, creates a keyword list, and controls and stores the keywords. Hand over to part 2401.

【0146】

The database 2403 used in this embodiment is the thesaurus data that represents the upper-lower relationship of the word concept. Figure 26 shows an example of thesaurus data. In FIG. 26, there is a superordinate / subordinate relationship such that action is a superordinate concept of Jurassic World and day hard and movie is a superordinate concept thereof. In addition, "day hard" is a proper noun representing a movie name, and "action" is a general noun representing a movie genre.

【0147】

Figure 27 shows an example of the processing procedure for creating a keyword list in the keyword extraction unit 2402.

【0148】

First, set the variable N to 1 (step S2902). This value represents the distance from the search string (supervisor in Figure 17) (step S2903). The distance mentioned here is a value of how many words are separated from the search character string.

【0149】

The morphological analysis result is determined whether or not it becomes a word-for-word keyword (steps S2903, S2904). The keywords referred to here are for narrowing down the search results, and refer to terms extracted by using part-speech information such as whether it is a noun or a proper noun, and field information such as computer terms and economic terms as keys. Here, the key for keyword extraction is a proper noun. The keyword words are stored in the keyword list (step S2911).

【0150】

In this specific example, when the morphological analysis result shown in FIG. 17 is input, "Jurassic World" is extracted as a keyword when the variable N = 2.

【0151】

Once the keywords have been extracted, the next step is to extract terms related to the keywords (steps S2912 to S2917).

【0152】

Here, a thesaurus is used as a database, and keyword candidates are extracted based on the distance between the keyword and the node on the thesaurus. The distance on the thesaurus here is the value of how many times the link is followed from the keyword node. That is, in FIG. 26, when viewed from the node of "Jurassic World", the node of "action" has a distance of 1, and the node of "day hard" and "movie" has a distance of 2.

【0153】

First, the variable M is set to 1 (step S2912). This value represents the distance on the thesaurus from the node indicating the keyword extracted in step S2904. It is determined whether the word on the thesaurus is a word-by-word keyword (steps S2913, S2914). The key for searching for keyword candidates using the thesaurus may be different from the key for searching for keyword candidates using the morphological analysis result. Here, the key is a general noun. The keyword words are stored in the keyword list (step S2915).

【0154】

In addition, determine if the variable M is out of scope (step S2916). The scope referred to here is a value indicating how far the distance on the thesaurus from the node representing the keyword is to be the target of keyword extraction. For example, if the scope is 5, keywords are extracted from nodes whose distance on the thesaurus from the node representing the keyword is 5 or less. In this embodiment, the scope is 2. If it does not exceed the scope, increment M by 1 (step S2917) and search for the keyword again. If so, all words in the scope in the thesaurus have been searched, so return to step S2906.

【0155】

In this specific example, when "Jurassic World" is extracted as a keyword, "action" and "movie" are extracted as nodes whose distance is 2 or less and which is a general noun from the keyword.

【0156】

In step S2906, it is determined whether the variable N is out of scope. The scope referred to here is a value indicating how many words away from the search character string are targeted for keyword extraction. For example, if the scope is 5, keywords will be extracted from the 5 words before and after the search string. In this embodiment, the scope is 3. If it does not exceed the scope, increase N by 1 in step S2907 and search for the keyword again. If it exceeds, all the words in the scope are searched, so the processing is terminated, and the keyword list that is the processing result is passed to the display content selection unit. If the keyword is not extracted, an empty keyword list is passed to the display content selection section.

【0157】

By the above processing, for example, when the morphological analysis result shown in FIG. 17 is input, "Jurassic World", "action", and "movie" are extracted as keywords when the variable N = 2, and the display content selection unit 2404. Passed to.

【0158】

The display content selection unit 2404 calculates an evaluation score for the meaning of the search result by using the method shown in the first embodiment, and passes the one having the higher score to the output unit 107. The keyword number table of this embodiment is shown in FIG. As a result, the second word meaning data is selected as the most appropriate word meaning.

【0159】

Then, for example, as shown in FIG. 29, the output unit 107 displays the dictionary search result.

【0160】

In the above, the thesaurus was used as the database 2403 in FIG. 25, but the thesaurus is not limited to this, and various databases can be used.

【0161】

Further, for example, in the first and second embodiments, when the input document is set to Japanese and the Japanese-English dictionary is used as the database 106 to be searched in FIG. 1, English data is obtained as the search result, and Japanese keywords are used. Even in cases where the search results cannot be selected, the data to be presented can be selected from the search results by converting the keywords into English translations using the Japanese-English dictionary as database 2403. .. Of course, the same is true for other languages.

【0162】

Further, for example, by converting a keyword into a translated word using a Japanese-English (English-Japanese) dictionary as the database 2403, the database search unit 104 can search the database using the translated word.

【0163】

Further, in the above, the superordinate concept of the keyword obtained by the sentence analysis was obtained by using the thesaurus (this is regarded as one conversion of the keyword). For example, first, the superordinate concept of the keyword was obtained by using the thesaurus. Next, a Japanese-English dictionary may be used to find a translation of the superordinate concept, and this may be used for a search or the like, so that the keyword may be converted two or more times.

【0164】

Further, for example, a plurality of different dictionaries may be used as the database 2403, and the keywords obtained by each dictionary may be used together.

【0165】

In addition, first, as in the first to fourth embodiments, the search process is performed using the keywords obtained from the sentence analysis results as they are, and when the keyword is not valid (for example, a word or phrase matching the keyword is found). If it does not exist in the search result or the headword), the search process may be performed using another one or more keywords obtained based on the keyword obtained from the sentence analysis result as in the present embodiment. ..

【0166】

(Sixth Embodiment) In the first to fifth embodiments, the case where the information is presented as the search result is character data has been described, but in the present embodiment, the first to fifth embodiments have been described. In the above, a case where a plurality of multimedia information such as a photograph, a figure, a voice, an image, a text, and the like are handled as the form will be described.

【0167】

When handling a plurality of multimedia information as a search result, not all multimedia information can be displayed (or played back, etc.) for an arbitrary search character string. For example, when the title of music appears in a sentence, audio data can be output, but a photograph or the like cannot be output because it is music.

【0168】

Therefore, when an arbitrary search character string is specified, it is preferable to present the type of information that can be displayed in the search character string to the user by using a pop-up menu or the like so that the user can select the information to be displayed. .. Some examples are shown below.

【0169】

Here, it is assumed that a document like the example in FIG. 16, that is, ".... He is the director of Jurassic World ..... ...." has already been entered.

【0170】

The user uses the input unit 101 to specify a character string (that is, a search character string) for which detailed information in the document is to be obtained.

【0171】

For example, if it is desired to obtain detailed information on the character string "supervisor", the user drags the character string portion "supervisor" using an input device such as a mouse, as illustrated in FIG. Then, in response to this, the search control unit presents displayable multimedia information to the user in a pop-up menu, as shown in FIG. 30, for example.

【0172】

The user then selects the type of detail he wants to get from the pop-up menu. For example, when it is desired to obtain detailed information on the "work" of the "director", the instruction is given by selecting the "work" part using an input device such as a mouse as shown in FIG.

【0173】

The information retrieval device uses a method as described above to display a list of "works" of "directors" instructed by the user, for example, as shown in FIG. 31. In addition, "movie A", "movie B", "movie C" and "movie D" are assumed to be proper nouns representing movie names.

【0174】

When "word meaning" is selected and instructed in the pop-up menu of FIG. 30, the word meaning of "director" is displayed as shown in FIG. 32, for example.

【0175】

The above designation may be performed by a function key, the number of mouse clicks, drag and drop of a designated icon, or the like.

【0176】

In addition, instead of searching for data after selecting the type of information as described above, information that can be displayed after searching for data in advance is presented, and search results corresponding to the type instructed by the user are displayed. You may try to do the same.

【0177】

Hereinafter, various variations in each of the above embodiments will be described.

【0178】

In the above, it has been described that data is passed to each unit via the search control unit under the control of the search control unit, but in data transfer, actual data may be passed, or instead. You may pass pointer information such as the address where the actual data is stored to.

【0179】

Further, in the above description, the database 106 is described as being provided in the information retrieval device, but the database 106 may not be provided in the information retrieval device, and the information retrieval device may access the external database 106. .. Further, in any case, a plurality of types of databases 106 may be searchable, and one or a plurality of databases to be searched may be specified by the user.

【0180】

Similarly, in the fifth embodiment, the database 2403 has been described as being provided in the search control unit 102, but the database 2403 is not provided in the information retrieval device, and the information retrieval device is an external database 2403. You may want to access.

【0181】

Further, when displaying the search result, the keyword used for the search or the like may be presented together.

【0182】

Further, the search functions of the above-described embodiments may be provided in an appropriate combination.

【0183】

Further, the functions described in each of the above embodiments can be realized by software, and the information retrieval device can be configured as a dedicated machine, but can also be configured by using a general-purpose computer. It is possible.

【0184】

Further, the functions described in each of the above embodiments are stored in a recording medium, that is, for causing the computer to perform a predetermined procedure (or for causing the computer to function as a predetermined means, or for the computer to perform a predetermined procedure). It can also be implemented as a computer-readable recording medium on which the program (to realize the function of) is recorded. For example, in FIG. 1, the search control unit 102, the sentence analysis unit 103, and the database search unit 104 can be implemented as a search control function, a sentence analysis function, and a database search function on a recording medium for information retrieval. Further, in this case, the analysis knowledge 105 or the database 106 may be stored in the recording medium to be realized, or may be stored in a storage medium (memory) different from the recording medium to perform a search control function. , The sentence analysis function, and the database search function may be made movable in cooperation with the recording medium in which the storage medium is stored.

【0185】

Further, the search function in each embodiment can be implemented by incorporating it into, for example, software for creating or editing a document, software for viewing a document acquired by WWW or e-mail, machine translation software, or the like. Is. It is also possible to implement the search service system as receiving document data and a designated search character string from an external request source via a network, for example, and transmitting the search result to the request source.

【0186】

The present invention is not limited to the above-described embodiment, and can be implemented with various modifications within the technical scope thereof.

【0187】

[Effect of the invention]

According to the present invention, the user can efficiently obtain necessary information, which is effective in accelerating the understanding of sentences. In other words, since the information is presented without the user narrowing down the necessary information by himself / herself, the load on the user is reduced.

[Simple explanation of drawings]

[Figure 1]

The figure which shows the structural example of the information retrieval apparatus which concerns on embodiment of this invention. [Figure 2]

Diagram showing a display example of a document [Fig. 3]

An example of the morphological analysis result of the target sentence is shown. [Fig. 4]

Diagram showing an example of Japanese dictionary search results [Fig. 5]

Diagram showing a configuration example of the search control unit [Fig. 6]

Flowchart showing an example of the procedure of keyword extraction processing using the morphological analysis result [Fig. 7]

Flowchart showing an example of processing procedure for selecting display information [Fig. 8]

A diagram showing an example of a keyword number table (initial state) [Fig. 9]

The figure which shows an example of the keyword number table (end state) [Fig. 10]

Diagram showing a display example of search results [Fig. 11]

Diagram showing a display example of search results [Fig. 12]

Diagram showing a display example of search results [Fig. 13]

Diagram showing a display example of search results [Fig. 14]

Diagram showing an example of the parsing result of the target sentence [Fig. 15]

Flowchart showing an example of the procedure of keyword extraction processing using the parsing result [Fig. 16]

Diagram showing a display example of a document [Fig. 17]

Diagram showing an example of the morphological analysis result of the target sentence [Fig. 18]

Diagram showing a configuration example of the search control unit [Fig. 19]

Flowchart showing an example of the procedure of keyword extraction processing using the morphological analysis result [Fig. 20]

Diagram showing an example of a search condition expression [Fig. 21]

Diagram showing an example of movie encyclopedia search results [Fig. 22]

Diagram showing a display example of search results [Fig. 23]

Diagram showing an example of the parsing result of the target sentence [Fig. 24]

Flowchart showing an example of the procedure of keyword extraction processing using the parsing result [Fig. 25]

Diagram showing a configuration example of the search control unit [Fig. 26]

Diagram showing an example of thesaurus data [Fig. 27]

Flowchart showing an example of processing procedure for selecting display information [Fig. 28]

The figure which shows an example of the keyword number table (end state) [Fig. 29]

Diagram showing a display example of search results [Fig. 30]

Diagram showing a display example of a pop-up menu [Fig. 31]

Diagram showing a display example of search results [Fig. 32]

Diagram showing a display example of search results [Explanation of symbols]

101 ... Input section 102 ... Search control unit 103 ... Sentence Analysis Department 104 ... Database Search Department 105 ... Analytical knowledge 106 ... database 107 ... Output 601 ... Control storage 602 ... Keyword extraction section 603 ... Display content selection section 2501 ... Control storage 2502 ... Search conditional expression creation department 2401 ... Control storage 2402 ... Keyword extraction section 2403 ... database 2404 ... Display content selection section