Document search system and document search method

Patent metadata

Field	Value
Publication number	US-12099543-B2
Application number	US-202017600280-A
Country	US
Kind code	B2
Filing date	Apr 16, 2020
Priority date	Apr 26, 2019
Publication date	Sep 24, 2024
Grant date	Sep 24, 2024

Abstract

Highly accurate document search, especially intellectual property-related document search, is achieved with a simple input method. A processing portion has a function of generating text analysis data from text data input to an input portion; a function of extracting a search word from words included in the text analysis data; and a function of generating first search data from the search word on the basis of weight dictionary data and thesaurus data. A memory portion stores second search data generated when the first search data is modified by a user. The processing portion updates the thesaurus data in accordance with the second search data.

First claim

The invention claimed is:

1. A document search system comprising an input portion, a database, a memory portion, a processing portion and an output portion,
wherein the database is configured to store a plurality of pieces of reference document data, weight dictionary data, and thesaurus data,
wherein the processing portion is configured to generate the weight dictionary data and the thesaurus data on the basis of the reference document data, generate text analysis data from text data input to the input portion, extract a search word from words included in the text analysis data, and generate first search data from the search word on the basis of the weight dictionary data and the thesaurus data,
wherein the output portion is configured to output the first search data to display on a GUI of a display device,
wherein the memory portion is configured to store second search data generated when the first search data displayed on the GUI is modified by a user, and
wherein the processing portion is configured to update the thesaurus data by adding a product of a contribution ratio and a difference between the first search data and the second search data to the thesaurus data.

2. The document search system according to claim 1 ,
wherein the processing portion is configured to generate reference text analysis data from the reference document data, and extract a plurality of keywords and related terms of the keywords from words included in the reference text analysis data.

3. The document search system according to claim 2 ,
wherein the weight dictionary data is generated by extracting appearance frequencies of the keywords from the words included in the reference text analysis data and adding, to each of the keywords, a first weight based on the appearance frequency.

4. The document search system according to claim 3 ,
wherein the first weight is a value based on an inverse document frequency of the keyword in the reference text analysis data.

5. The document search system according to claim 3 ,
wherein the thesaurus data is generated by adding a second weight to each of the related terms.

6. The document search system according to claim 5 ,
wherein the second weight is a product of the first weight of the keyword and a value based on a similarity degree or a distance between a distributed representation vector of the related term and a distributed representation vector of the keyword.

7. The document search system according to claim 6 ,
wherein the distributed representation vector is generated with use of a neural network.

8. The document search system according to claim 1 ,
wherein the processing portion comprises a transistor, and
wherein the transistor comprises a metal oxide in its channel formation region.

9. The document search system according to claim 1 ,
wherein the processing portion comprises a transistor, and
wherein the transistor comprises silicon in its channel formation region.

10. The document search method according to claim 1 ,
wherein, in updating the thesaurus data, the processing portion is configured to update on the basis of second search data modified by a plurality of users.

11. A document search method comprising the steps of:
generating weight dictionary data and thesaurus data on the basis of a plurality of pieces of reference document data;
generating text analysis data from text data;
extracting a search word from words included in the text analysis data;
generating first search data from the search word on the basis of the weight dictionary data and the thesaurus data;
outputting the first search data to display on a GUI of a display device;
generating second search data when the first search data displayed on the GUI is modified by a user;
updating the thesaurus data by adding a product of a contribution ratio and a difference between the first search data and second search data; and
generating ranking data by giving scores to the plurality of pieces of reference document data on the basis of the second search data and ranking the plurality of pieces of reference document data on the basis of the scores to the thesaurus data.

12. The document search method according to claim 11 ,
wherein reference text analysis data is generated from the reference document data, and
wherein a plurality of keywords and related terms of the keywords are extracted from words included in the reference text analysis data.

13. The document search method according to claim 12 ,
wherein the weight dictionary data is generated by extracting appearance frequencies of the keywords from the words included in the reference text analysis data and adding, to each of the plurality of keywords, a first weight based on the appearance frequency.

14. The document search method according to claim 13 ,
wherein the first weight is a value based on an inverse document frequency of the keyword in the reference text analysis data.

15. The document search method according to claim 14 ,
wherein the thesaurus data is generated by adding a second weight to each of the related terms.

16. The document search method according to claim 15 ,
wherein the second weight is a product of the first weight of the keyword and a value based on a similarity degree or a distance between a distributed representation vector of the related term and a distributed representation vector of the keyword.

17. The document search method according to claim 16 ,
wherein the distributed representation vector is generated with use of a neural network.

18. The document search method according to claim 11 ,
wherein, in updating the thesaurus data, the second search data is modified by a plurality of users.

Assignees

• Semiconductor Energy Lab

Inventors

• Higashi Kazuki
• Momo Junpei

Classifications

• G06N3/09

  Supervised learning · CPC title

• G06N3/0499

  Feedforward networks · CPC title

• G06F40/242

  Dictionaries · CPC title

• G06F40/247

  Thesauruses; Synonyms · CPC title

• G06F40/279

  Recognition of textual entities · CPC title