System and method of making search for document in accordance with query of natural language

• US 7,765,201 B2
• Filed: 03/14/2006
• Issued: 07/27/2010
• Est. Priority Date: 03/14/2005
• Status: Expired due to Fees

• Current Assignee: Kabushiki Kaisha Toshiba (Toshiba Corporation), Toshiba Solutions Corporation (Toshiba Corporation)
• Original Assignee: Kabushiki Kaisha Toshiba (Toshiba Corporation), Toshiba Solutions Corporation (Toshiba Corporation)
• Inventors: Sasaki, Koichi, Takachio, Katsuhiko, Sasaki, legal representative
• Primary Examiner(s): Mahmoudi; Tony
• Assistant Examiner(s): Hu; Jensen
• Application Number:
  US11/373,991
• Publication Number:
  US 20060206463A1
• Time in Patent Office: 1,596 Days
• Field of Search: 707/4
• US Class Current: 707/713
• CPC Class Codes: G06F 16/3329 Natural language query formulation or dialogue systems

Abstract

A first interface accepts a user specified first question sentence. A question analysis unit analyzes the first question sentence. When the first question sentence queries a plurality of matters, the question analysis unit generates a plurality of second question sentences for individually querying each of the plurality of matters from the first question sentence. A search engine makes searches for documents which match respective matters queried by the plurality of second question sentences from a morphological index database in accordance with the plurality of second question sentences. A second interface provides a user with a search result for the first question having reflected thereon results of document searches for the plurality of second question sentences by the search engine.

Claims

1. A apparatus for making a search for a document in accordance with a query of a natural language, comprising:

a first interface unit configured to input a user specified first question sentence represented in a natural language, the first question sentence including a term representing a query;

a morphological analysis unit configured to execute morphological analysis for the first question sentence input by the first interface unit, thereby to divide the question sentence on a word by word basis;

a question analysis unit configured to analyze the first question sentence, thereby to generate a plurality of second question sentences from the first question sentence, wherein the question analysis unit includes;

a first module configured to specify, based on a morphological analysis result by the morphological analysis unit, a first noun that corresponds to a subjective case of the first question sentence and is included in the first question sentence;

a second module configured to extract one second noun or a plurality of second nouns, which are included in the first question sentence, other than the first noun from the first question sentence based on the morphological analysis result; and

a third module configured to connect the first noun, each of the plurality of second nouns, and the term representing the query included in the first question sentence for said each of the plurality of second nouns when the plurality of second nouns are extracted, thereby to generate the plurality of second question sentences, each including the first noun, at least one of the plurality of second nouns, and the term representing the query, for individually querying each of a plurality of matters, the first noun also corresponding to the subjective case of each of the plurality of second question sentences;

a search engine configured to make searches for documents which match respective matters queried by the plurality of second question sentences from a morphological index database by index searches according to the plurality of second question sentences generated by the question analysis unit, the morphological index database storing morphological analysis results for a plurality of documents as indexes;

a storage unit which stores tuning information, the tuning information including a concatenation condition for specifying which of documents acquired by the document searches by the search engine and abstracts of the acquired documents are used as search results for the plurality of second question sentences;

a concatenation unit configured to select the documents acquired by the document searches by the search engine or the abstracts of the acquired documents as the search results for the plurality of second question sentences to be concatenated with each other in accordance with the tuning information stored in the storage unit, thereby to generate a search result document which represents a search result for the first question sentence by concatenating search results for the plurality of second question sentences by the search engine; and

a second interface unit configured to provide a user with the search result document generated by the concatenation unit as a search result for the first question sentence.

2. The apparatus according to claim 1, wherein, when a search result for one of the plurality of second question sentences includes a plurality of search results analogous to each other, the concatenation unit handles only one of the analogous search results as a search result for the one second question sentence.

3. The apparatus according to claim 1, further comprising an abstract generation unit configured to generate abstracts of documents acquired by the document searches for the plurality of second question sentences by the search engine.

4. The apparatus according to claim 1, further comprising a third interface unit configured to provide a user with a user interface screen for causing the user to specify the concatenation condition, the third interface unit storing in the storage unit the tuning information including the concatenation condition specified by the user on the user interface screen.

5. The apparatus according to claim 4, wherein the user interface screen includes an area which causes a user to specify which of the documents acquired by the document searches by the search engine and the abstracts of the acquired documents are used as search results for the plurality of second question sentences to be concatenated with each other by the concatenation unit.

6. The apparatus according to claim 4, wherein:

the user interface screen includes an area which causes a user to specify a maximum number of lines of a document which is usable for concatenation by the concatenation unit, and

when the number of lines of a document to be concatenated exceeds the maximum number of lines specified by the user on the user interface screen, the concatenation unit uses an abstract of the document instead of the document to be concatenated.

7. The apparatus according to claim 4, wherein, when a plurality of search results exist with respect to one of the plurality of second question sentences, the user interface screen includes an area which causes the user to instruct the apparatus whether to concatenate the plurality of search results with each other as a search result for the one second question sentence.

8. The apparatus according to claim 1, wherein the concatenation unit sets link information for referring documents acquired by document searches for the plurality of second question sentences by the search engine to a search result document which represents a search result for the first question sentence.

9. A method for making a search for a document in accordance with a query of a natural language, in document search apparatus with a storage unit, comprising:

inputting a user specified first question sentence represented in a natural language, the first question sentence including a term representing a query;

executing morphological analysis for the first question sentence, thereby to divide the question sentence on a word by word basis;

analyzing the first question sentence, thereby to generate a plurality of second question sentences from the first question sentence, wherein the analyzing includes;

specifying, based on a morphological analysis result by the morphological analysis, a first noun that corresponds to a subjective case of the first question sentence and is included in the first question sentence;

extracting one second noun or a plurality of second nouns, which are included in the first question sentence, other than the first noun from the first question sentence based on the morphological analysis result; and

connecting the first noun, each of the plurality of second nouns, and the term representing the query included in the first question sentence for said each of the plurality of second nouns when the plurality of second nouns are extracted, thereby to generate the plurality of second question sentences, each including the first noun, at least one of the plurality of second nouns, and the term representing the query, for individually querying each of a plurality of matters, the first noun also corresponding to the subjective case of each of the plurality of second question sentences;

making searches for documents which match respective matters queried by the plurality of second question sentences from a morphological index database by index searches according to the plurality of second question sentences, the morphological index database storing morphological analysis results for a plurality of documents as indexes;

selecting the documents acquired by the document searches or the abstracts of the acquired documents as the search results for the plurality of second question sentences to be concatenated with each other in accordance with tuning information stored in the storage unit, the tuning information including a concatenation condition for specifying which of documents acquired by the document searches and abstracts of the acquired documents are used as search results for the plurality of second question sentences;

generating a search result document which represents a search result for the first question sentence by concatenating search results for the plurality of second question sentences by the document searches; and

providing a user with the search result document as a search result for the first question sentence.

10. A computer-readable storage medium storing a computer program product for use in a document search in accordance with a query of a natural language, the computer program product being employed in a computer with a storage unit, the computer program product implementing a method comprising:

causing the computer to input a user specified first question sentence represented in a natural language, the first question sentence including a term representing a query;

causing the computer to execute morphological analysis for the first question sentence, thereby to divide the question sentence on a word by word basis;

causing the computer to analyze for the first question sentence, thereby to generate a plurality of second question sentences from the first question sentence, wherein the causing the computer to analyze includes;

causing the computer to specify, based on a morphological analysis result by the morphological analysis, a first noun that corresponds to a subjective case of the first question sentence and is included in the first question sentence;

causing the computer to extract one second noun or a plurality of second nouns other than the first noun, which are included in the first question sentence, from the first question sentence based on the morphological analysis result; and

causing the computer to connect the first noun, each of the plurality of second nouns, and the term representing the query included in the first question sentence for said each of the plurality of second nouns when the plurality of second nouns are extracted, thereby to generate the plurality of second question sentences, each including the first noun, at least one of the plurality of second nouns, and the term representing the query, for individually querying each of a plurality of matters, the first noun also corresponding to the subjective case of each of the plurality of second question sentences;

causing the computer to search for documents which match respective matters queried by the plurality of second question sentences from a morphological index database by executing index searches in accordance with the plurality of second question sentences, the morphological index database storing morphological analysis results for a plurality of documents as indexes;

causing the computer to select the documents acquired by the document searches or the abstracts of the acquired documents as the search results for the plurality of second question sentences to be concatenated with each other in accordance with tuning information stored in the storage unit, the tuning information including a concatenation condition for specifying which of documents acquired by the document searches and abstracts of the acquired documents are used as search results for the plurality of second question sentences;

causing the computer to generate a search result document which represents a search result for the first question sentence by concatenating search results for the plurality of second question sentences by the document searches; and

causing the computer to provide a user with the search result document as a search result for the first question sentence.