Ranking entity search results based on information density

Patent metadata

Field	Value
Publication number	US-12045243-B2
Application number	US-202117542412-A
Country	US
Kind code	B2
Filing date	Dec 4, 2021
Priority date	Dec 4, 2021
Publication date	Jul 23, 2024
Grant date	Jul 23, 2024

Abstract

The method provides for receiving a plurality of documents including mentions of a target entity from a search query about the entity. The mentions of the target entity are identified in respective documents of the plurality of documents. Content surrounding the one or more mentions of the target entity are extracted with the mentions within the respective documents and form section. A respective document includes a plurality of sections. Metrics of relevance and irrelevance to the target entity are determined within the plurality of sections of the respective documents. A density score is generated for the plurality of sections of the respective documents. A relevancy score is assigned to respective documents of the plurality of documents, based on the density scores of the sections of the respective documents. The documents are ranked based on the relevancy score and presented in an order based on the ranking.

First claim

What is claimed is:

1. A method for ranking documents of a search result according to relevance of a query of an entity, the method comprising:
receiving a plurality of documents in which respective documents include one or more mentions of a target entity resulting from a search query about the target entity;
identifying the one or more mentions of the target entity in the respective documents of the plurality of documents;
extracting content surrounding the one or more mentions of the target entity within the respective documents and including the one or more mentions of the target entity and the extracted content as a section of a plurality of sections of a respective document of the plurality of documents;
determining metrics of relevance and irrelevance of other entities mentioned within the extracted content surrounding the one or mentions of the target entity, within the plurality of sections of the respective documents, to the target entity, wherein a relevant entity shares a relationship with the target entity if the relevant entity and the target entity are mentioned in a same section of extracted content among a pre-determined number of other respective documents;
generating an information density score, for the plurality of sections of the respective documents, based on a combination of a number of relevant entities and an identifier entity count with respect to the number of irrelevant entities, wherein irrelevant other entities lack a relationship with the target entity that is supported by a pre-determined number of other documents of the plurality of documents resulting from the search query, and wherein an identifier entity includes information associated with the identity of the target entity;
assigning a relevancy score to the respective documents of the plurality of documents, based on the density scores of the plurality of sections of the respective documents, wherein the relevancy score equals a highest density score of a section of the document;
ranking the respective documents based on the relevancy score; and
presenting the respective documents in an order based on the ranking.

2. The method according to claim 1 , wherein the entity comprises a person.

3. The method according to claim 1 , wherein the extracted content comprises a sentence that includes the mention of the target entity and a set of sentences surrounding the mention in the document.

4. The method according to claim 1 , further comprising; creating a sub-corpus of distinct document classes including a selection of a subset of the plurality of documents, based on applying a similarity metric to the plurality of documents.

5. The method of claim 1 , wherein identifying the mention includes identifying a sentence in which the mention occurs within a document of the plurality of documents and referring to the sentence as a snippet of the document; and
combining two or more snippets in which content extracted that surrounds the mention of the target entity overlaps between the two or more snippets, and the combined snippets and corresponding content extracted, resulting in a section of the document.

6. A computer program product for ranking documents of a search result according to relevance of a query of an entity, the computer program product comprising:
one or more computer-readable storage media, and program instructions stored on the one or more computer-readable storage media, the program instructions comprising:
program instructions to receive a plurality of documents in which respective documents include one or more mentions of a target entity resulting from a search query about the target entity;
program instructions to identify the one or more mentions of the target entity in the respective documents of the plurality of documents;
program instructions to extract content surrounding the one or more mentions of the target entity within the respective documents and including the one or more mentions of the target entity and the extracted content as a section of a plurality of sections of a respective document of the plurality of documents;
program instructions to determine metrics of relevance and irrelevance of other entities mentioned within the extracted content surrounding the one or mentions of the target entity, within the plurality of sections of the respective documents, to the target entity, wherein a relevant entity shares a relationship with the target entity if the relevant entity and the target entity are mentioned in a same section of the extracted content among a pre-determined number of other respective documents;
program instructions to generate an information density score, for the plurality of sections of the respective documents, based on a combination of a number of relevant entities and an identifier entity count with respect to the number of irrelevant entities, wherein irrelevant other entities lack a relationship with the target entity that is supported by a pre-determined number of other documents of the plurality of documents resulting from the search query, and wherein an identifier entity includes information associated with the identity of the target entity;
program instructions to assign a relevancy score to the respective documents of the plurality of documents, based on the density scores of the plurality of sections of the respective documents, wherein the relevancy score equals a highest density score of a section of the document;
program instructions to ranking the respective documents based on the relevancy score; and
program instructions to present the respective documents in an order based on the ranking.

7. The computer program product of claim 6 , further comprising:
program instructions to create a sub-corpus of distinct document classes including a selection of a subset of the plurality of documents, based on applying a similarity metric to the plurality of documents.

8. The computer program product of claim 6 , further comprising:
program instructions to identify a sentence in which the mention occurs within a document of the plurality of documents and referring to the sentence as a snippet of the document; and
program instructions to combine two or more snippets in which content extracted that surrounds the mention of the target entity overlaps between the two or more snippets, and the combined snippets and corresponding content extracted, resulting in a resulting in a section of the document.

9. A computer system for ranking documents of a search result according to relevance of a query of an entity, the computer program product comprising:
one or more computer processors;
one or more computer-readable storage media, and program instructions stored on the one or more computer-readable storage media, the program instructions comprising:
program instructions to receive a plurality of documents in which respective documents include one or more mentions of a target entity resulting from a search query about the target entity;
program instructions to identify the one or more mentions of the target entity in respective documents of the plurality of documents;
program instructions to extract content surrounding the one or more mentions of the target entity within the respective documents and including the one or more mentions of the target entity and the extracted content as a section of a plurality of sections of a respective document of the plurality of documents;
program instructions to determine metrics of relevance and irrelevance of other entities mentioned within the extracted content surrounding the one or mentions of the target entity, within the plurality of sections of the respective documents, to the target entity, wherein a relevant entity shares a relationship with the tar

Assignees

• IBM

Inventors

• Ackermann Christopher F
• Beller Charles E
• Drzewucki Michael

Classifications

• G06F16/93

  Document management systems · CPC title

• G06F16/248

  Presentation of query results · CPC title

• G06F16/24578 Primary

  using ranking · CPC title