Information search support method and information search support device

• US 8,099,418 B2
• Filed: 05/14/2008
• Issued: 01/17/2012
• Est. Priority Date: 05/28/2007
• Status: Expired due to Fees

• Current Assignee: Panasonic Intellectual Property Corporation of America (Panasonic Holdings Corporation)
• Original Assignee: Panasonic Corporation (Panasonic Holdings Corporation)
• Inventors: Inoue, Tsuyoshi, Matsuura, Satoshi
• Primary Examiner(s): Wu, Yicun
• Application Number:
  US12/517,597
• Publication Number:
  US 20100281036A1
• Time in Patent Office: 1,343 Days
• Field of Search: 707/748, 707/754, 707/769
• US Class Current: 707/748
• CPC Class Codes: G06F 16/90335 Query processing

Abstract

An information search support device includes: a cluster selection history information accumulation unit which accumulates content information of each cluster and cluster selection history information indicating a cluster selected by a user; a degree of certainty calculation unit which calculates a degree of certainty indicating a degree to which a vague search target of the user has been identified, based on the cluster selection history information accumulated; a search target estimation unit which obtains a condition for estimating the search target; a search-omitted information extraction unit which extracts search-omitted information that is included in the cluster that is not selected and is estimated from the obtained condition; and a search-omitted information addition unit which adds, to the cluster selected by the user, the extracted search-omitted information, in the case where the degree of certainty is equal to or higher than a predetermined threshold.

Claims

1. An information search support device which supports information search by a user, said information search support device comprising:

a display unit configured to perform a first clustering on search-target information, and to display a plurality of clusters and information indicating a feature of each of the displayed clusters;

a clustering unit configured to perform a second clustering on a search target of the user, the search target of the user including information included in a cluster selected by the user from among the displayed clusters;

a processor;

a cluster selection history information accumulation unit configured to accumulate, using said processor, cluster selection history information as content information of each of the displayed clusters at a time of performing the second clustering, the cluster selection history information indicating one of the following;

a number of pieces of information included in the cluster selected by the user;

a number of clusters selected from among the displayed clusters, and a number of documents including a representative word representing the cluster selected by the user among documents indicating contents of the search-target information;

a degree of certainty calculation unit configured to calculate a degree of certainty indicating a degree of clarity for the search target of the user, based on the cluster selection history information accumulated in said cluster selection history information accumulation unit;

a search target estimation unit configured to obtain, from content information of the cluster selected by the user, a condition for estimating the search target of the user;

a search-omitted information extraction unit configured to extract, from a cluster that is not selected by the user from among the displayed clusters, search-omitted information that is (i) information included in the cluster that is not selected by the user and (ii) information belonging to a set of information estimated from the condition obtained by said search target estimation unit; and

a search-omitted information addition unit configured to (i) judge whether or not the degree of certainty calculated by said degree of certainty calculation unit is equal to or higher than a predetermined threshold, and (ii) add, to the cluster selected by the user, the search-omitted information extracted by said search-omitted information extraction unit, in a case where the judgment shows that the degree of certainty calculated by said degree of certainty calculation unit is equal to or higher than the predetermined threshold,

wherein said clustering unit is configured to perform a third clustering on the search target of the user after said search-omitted information addition unit adds the search-omitted information to the cluster selected by the user.

2. The information search support device according to claim 1,

wherein said search target estimation unit is configured to obtain, as the condition for estimating the search target of the user, a representative word representing the cluster selected by the user, and

wherein said search-omitted information extraction unit is configured to generate an extraction vector for extracting the search-omitted information from the information included in the cluster that is not selected by the user, using a vector obtained from the representative word obtained by said search target estimation unit.

3. The information search support device according to claim 1,

wherein said cluster selection history information accumulation unit is configured to accumulate, as the content information of each of the displayed clusters, one of the following:

the number of pieces of information included in the cluster selected by the user;

the number of clusters selected from among the displayed clusters; and

the number of documents including a representative word representing the cluster selected by the user among documents indicating contents of the search-target information, and

wherein said degree of certainty calculation unit is configured to calculate, for certain content information accumulated in said cluster selection history information accumulation unit, the degree of certainty according to a rule that the degree of certainty becomes higher as the number of pieces of content information becomes smaller.

4. The information search support device according to claim 1,

wherein said degree of certainty calculation unit is configured to calculate a degree of semantic similarity between a representative word representing a previously selected cluster and a representative word representing a currently selected cluster, and to calculate the degree of certainty according to a mathematical expression that the degree of certainty becomes higher as the calculated degree of semantic similarity becomes higher.

5. The information search support device according to claim 1,

wherein said search-omitted information extraction unit is configured to (i) determine a degree of importance for a representative word to be used for extracting the search-omitted information, (ii) determine, with the degree of importance, that the representative word becomes more important as the number of documents in which the representative word appears becomes smaller, the documents indicating contents of the search-target information, and (iii) generate an extraction vector from the representative word after the representative word is weighted according to the determined degree of importance.

6. The information search support device according to claim 1,

wherein said search-omitted information extraction unit is configured to (i) calculate a degree of semantic similarity between a representative word representing a previously selected cluster and a representative word representing a currently selected cluster, (ii) determine that a consistency between (a) the representative word of the previously selected cluster and (b) the representative word of the currently selected cluster becomes stronger as the calculated degree of semantic similarity becomes higher, (iii) determine, for a representative word to be used for extracting the search-omitted information, a degree of importance so that the degree of importance of the representative word to be used for extracting the search-omitted information becomes higher as the consistency between (a) the representative word of the previously selected cluster and (b) the representative word of the currently selected cluster becomes stronger, and (iv) generate an extraction vector from the representative word to be used for extracting the search-omitted information after the representative word to be used for extracting the search-omitted information is weighted according to the determined degree of importance.

7. An information search support method for supporting information search by a user, wherein said information search support method is performed using a processor, said information search support method comprising:

(i) performing a first clustering on search-target information, and displaying a plurality of clusters and information indicating a feature of each of the displayed clusters, said performing the first clustering and displaying being performed by a display unit;

(ii) performing a second clustering on a search target of the user, the search target of the user including information included in a cluster selected by the user from among the displayed clusters, said performing the second clustering being performed by a clustering unit;

(iii) accumulating, using the processor, cluster selection history information as content information of each of the displayed clusters at a time of performing the second clustering, the cluster selection history information indicating one of the following;

a number of pieces of information included in the cluster selected by the user;

a number of clusters selected from among the displayed clusters, and a number of documents including a representative word representing the cluster selected by the user among documents indicating contents of the search-target information, said accumulating being performed by a cluster selection history information accumulation unit;

(iv) calculating a degree of certainty indicating a degree of clarity for the search target of the user, based on the cluster selection history information accumulated in said accumulating;

(v) obtaining, from content information of the cluster selected by the user, a condition for estimating the search target of the user;

(vi) extracting, from a cluster that is not selected by the user from among the displayed clusters, search-omitted information that is (a) information included in the cluster that is not selected by the user and (b) information belonging to a set of information estimated from the condition obtained in said obtaining;

(vii) judging whether or not the degree of certainty calculated in said calculating is equal to or higher than a predetermined threshold;

(viii) adding, to the cluster selected by the user, the search-omitted information extracted in said extracting, in a case where the judgment shows that the degree of certainty is equal to or higher than the predetermined threshold; and

(ix) performing a third clustering on the search target of the user after said adding adds the search-omitted information to the cluster selected by the user.

10. The information search support method according to claim 7, further comprising:

displaying information indicating a feature of each of one or more clusters on which the third clustering is performed, said displaying being performed by the display unit; and

performing a fourth clustering on the search target of the user, the search target of the user including a cluster selected again by the user from among the displayed clusters, said performing the fourth clustering being performed by the clustering unit.

8. A non-transitory computer-readable recording medium for use in a computer, the recording medium having a computer program recorded thereon for causing the computer to execute an information search support method, the information search support method comprising:

performing a first clustering on search-target information, and displaying a plurality of clusters and information indicating a feature of each of the displayed clusters, said performing the first clustering and displaying being performed by a display unit;

performing a second clustering on a search target of a user, the search target of the user including information included in a cluster selected by the user from among the displayed clusters, said performing the second clustering being performed by a clustering unit;

accumulating cluster selection history information as content information of each of the displayed clusters at a time of performing the second clustering, the cluster selection history information indicating one of the following;

a number of pieces of information included in the cluster selected by the user;

a number of clusters selected from among the displayed clusters, and a number of documents including a representative word representing the cluster selected by the user among documents indicating contents of the search-target information, said accumulating being performed by a cluster selection history information accumulation unit;

calculating a degree of certainty indicating a degree of clarity for the search target of the user, based on the cluster selection history information accumulated in said accumulating;

obtaining, from content information of the cluster selected by the user, a condition for estimating the search target of the user;

extracting, from a cluster that is not selected by the user from among the displayed clusters, search-omitted information that is (i) information included in the cluster that is not selected by the user and (ii) information belonging to a set of information estimated from the condition obtained in said obtaining;

judging whether or not the degree of certainty calculated in said calculating is equal to or higher than a predetermined threshold;

adding, to the cluster selected by the user, the search-omitted information extracted in said extracting, in a case where the judgment shows that the degree of certainty is equal to or higher than the predetermined threshold; and

performing a third clustering on the search target of the user after said adding adds the search-omitted information to the cluster selected by the user.

9. An information search support device which supports information search by a user, said information search support device comprising:

a display unit configured to perform a first clustering on search-target information, and to display a plurality of clusters and information indicating a feature of each of the displayed clusters;

a clustering unit configured to perform a second clustering on a search target of the user, the search target of the user including information included in a cluster selected by the user from among the displayed clusters;

a processor;

a cluster selection history information accumulation unit configured to accumulate, using said processor, cluster selection history information as content information of each of the displayed clusters at a time of performing the second clustering, the cluster selection history information indicating one of the following;

a number of pieces of information included in the cluster selected by the user;

a number of clusters selected from among the displayed clusters, and a number of documents including a representative word representing the cluster selected by the user among documents indicating contents of the search-target information;

a degree of certainty calculation unit configured to calculate a degree of certainty indicating a degree of clarity for the search target of the user, based on the cluster selection history information accumulated in said cluster selection history information accumulation unit;

a search target estimation unit configured to obtain, from content information of the cluster selected by the user, a condition for estimating the search target of the user; and

a search-omitted information extraction unit configured to extract, from a cluster that is not selected by the user from among the displayed clusters, search-omitted information that is (i) information included in the cluster that is not selected by the user and (ii) information belonging to a set of information estimated from the condition obtained by said search target estimation unit.