Search apparatus, search program, and search method

• US 7,734,614 B2
• Filed: 05/25/2007
• Issued: 06/08/2010
• Est. Priority Date: 06/09/2006
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Aoki, Kazuo, Aramaki, Kazuhiro
• Primary Examiner(s): Trujillo; James
• Assistant Examiner(s): Mamillapalli; Pavan
• Application Number:
  US11/753,769
• Publication Number:
  US 20080059431A1
• Time in Patent Office: 1,110 Days
• Field of Search: 706/13, 715/531, 707/4, 707/101, 707/5, 707/3
• US Class Current: 707/711
• CPC Class Codes: G06F 16/319 Inverted lists

Abstract

The search apparatus of the present invention searches for a document corresponding to a specified search term from among a plurality of documents, and includes a search term pair generation unit for generating at least one search term pair including a first search string with a length equal to the number of index characters and a second search string with a length equal to the number of index characters located at a position shifted by a predetermined number of offset characters relative to the first search string, which are included in the search term, a search unit for searching, for each search term pair, an index database for a document which has both of the first search string and the second search string as indexing terms, a calculation unit for calculating a score of each document on the basis of a frequency of occurrence of the first search string and the second search string of each search term pair included in each document, and a selection unit for selecting a document to be outputted as a search result from among the plurality of documents, on the basis of the respective scores of the plurality of documents.

Claims

1. A search apparatus for searching for a document corresponding to a specified search term from among a plurality of documents, the search apparatus comprising:

an indexing term storage unit for storing an index database having a plurality of character strings each of which is composed of successive characters in each of a plurality of documents with a length equal to a predetermined number of index characters, and is used as an indexing term for that document;

a search term pair generation unit for generating at least one search term pair including a first search string with a length equal to the number of index characters and a second search string with a length equal to the number of index characters located at a position shifted by a predetermined number of offset characters relative to the first search string, which are included in the search term;

a search unit for searching the index database for a document having both of the first search string and the second search string as indexing terms, for each of the search term pairs;

a calculation unit for calculating a score indicating a degree at which each document is included in a search result on the basis of a frequency of occurrence of the first search string and the second search string of each of the search term pairs included in each document or of whether or not the first search string and the second search string are included in each document;

a selection unit for selecting a document to be outputted as the search result from among the plurality of documents, on the basis of the respective scores of the plurality of documents; and

an output unit for outputting the document selected by the selection unit as the search result.

2. The search apparatus according to claim 1, wherein the indexing term storage unit stores, in association with each of the indexing terms, each document including that indexing term and a partial score corresponding to that indexing term in each document in the index database, and the calculation unit calculates, for each document, the score of that document on the basis of the partial scores corresponding to the first search string and the second search string of each of the search term pairs.

3. The search apparatus according to claim 2, wherein the calculation unit calculates the score of the document on the basis of a smaller one of the partial scores corresponding to the first search string and the second search string for each of the search term pairs.

4. The search apparatus according to claim 1, wherein the indexing term storage unit stores, in association with to each of the indexing terms, each document including that indexing term and a position of that indexing term in each document in the index database, and the search unit extracts a document for the search term pair, on condition that the second search string is included at a position shifted by the number of offset characters relative to the first search string included in the document.

5. A search apparatus for searching for a document corresponding to a specified search term from among a plurality of documents, the search apparatus comprising:

an indexing term storage unit for storing an index database having a plurality of character strings each of which is composed of successive characters in each of a plurality of documents with a length equal to a predetermined number of index characters, and is used as an indexing term for that document;

a generation unit for generating at least one search string with a length equal to the number of index characters, the search string being included in the search term;

a partial search unit for searching the index database for a document which has each of the search strings as an indexing term;

a calculation unit for calculating a score indicating a degree at which each document is included in a search result, on the basis of a frequency of occurrence of each of the search strings included in each document or of whether or not each of the search strings is included in each document;

a threshold generation unit for generating a threshold of the score which becomes larger as the search term becomes longer, and which becomes smaller as the search term becomes shorter;

a selection unit for selecting a document having the score higher than the threshold as a document to be outputted as the search result, from among the plurality of documents; and

an output unit for outfitting the document selected by the selection unit.

6. The search apparatus according to claim 5, further comprising an exact match search unit for searching for a document including the search term from among the plurality of documents, wherein

the calculation unit further calculates the score of the document, on the basis of a frequency of occurrence of the search term included in the document, and

the threshold generation unit generates the threshold on the basis of the score of the document including the search term.

7. The search apparatus according to claim 6, wherein the threshold generation unit generates the threshold smaller than a minimum value of the scores of all the documents including the search term.

8. The search apparatus according to claim 7, wherein the threshold generation unit generates the threshold using a function which includes a linear expression of the length of the search term as a denominator and gradually increases toward the minimum value as the length of the search term becomes longer.

9. A search program for searching for a document corresponding to a specified search term from among a plurality of documents by a computer, the search program causing the computer to function as:

an indexing term storage unit for storing an index database having a plurality of character strings each of which is composed of successive characters in each of a plurality of documents with a length equal to a predetermined number of index characters, and is used as an indexing term for that document;

a search term pair generation unit for generating at least one search term pair including a first search string with a length equal to the number of index characters and a second search string with a length equal to the number of index characters located at a position shifted by a predetermined number of offset characters relative to the first search string, which are included in the search term;

a search unit for searching the index database for a document having both of the first search string and the second search string as indexing terms, for each of the search term pairs;

a calculation unit for calculating a score indicating a degree at which each document is included in a search result on the basis of a frequency of occurrence of the first search string and the second search string of each of the search term pairs included in each document or of whether or not the first search string and the second search string are included in each document;

a selection unit for selecting a document to be outputted as the search result from among the plurality of documents, on the basis of the respective scores of the plurality of documents; and

an output unit for outputting the document selected by the selection unit as the search result.

10. A search method for searching for a document corresponding to a specified search term from among a plurality of documents by a computer, wherein the computer stores an index database having a plurality of character strings each of which is composed of successive characters in each of a plurality of documents with a length equal to a predetermined number of index characters, and is used as an indexing term for that document, the method comprising:

a search term pair generation step of generating, by the computer, at least one search term pair including a first search string with a length equal to the number of index characters and a second search string with a length equal to the number of index characters located at a position shifted by a predetermined number of offset characters relative to the first search string, which are included in the search term;

a search step of searching, by the computer, the index database for a document having both of the first search string and the second search string as indexing terms, for each of the search term pairs;

a calculation step of calculating, by the computer, a score indicating a degree at which each document is included in a search result on the basis of a frequency of occurrence of the first search string and the second search string of each of the search term pairs included in each document or of whether or not the first search string and the second search string are included in each document;

a selection step of selecting, by the computer, a document to be outputted as the search result, on the basis of the respective scores of the plurality of documents; and

an output step of outputting the document selected by the selection step.