Document search system, document search method, program, and non-transitory computer readable storage medium

Patent metadata

Field	Value
Publication number	US-11789953-B2
Application number	US-201916979197-A
Country	US
Kind code	B2
Filing date	Mar 13, 2019
Priority date	Mar 23, 2018
Publication date	Oct 17, 2023
Grant date	Oct 17, 2023

Abstract

A highly accurate document search, particularly a search for a document relating to intellectual property, is achieved with an easy input method. A document search system includes a processing portion. The processing portion has a function of extracting a keyword included in text data, a function of extracting a related term of the keyword from words included in a plurality of pieces of first reference text analysis data, a function of giving a weight to each of the keyword and the related term, a function of giving a score to each of a plurality of pieces of second reference text analysis data on the basis of the weight, a function of ranking the plurality of pieces of second reference text analysis data on the basis of the score to generate ranking data, and a function of outputting the ranking data.

First claim

The invention claimed is:

1. A document search system comprising:
a processing portion,
wherein the processing portion is configured to extract a keyword included in text data, to extract a related term of the keyword from words included in a plurality of pieces of first reference text analysis data, to give a weight to each of the keyword and the related term, to give a score to each of a plurality of pieces of second reference text analysis data on the basis of the weight, to rank the plurality of pieces of second reference text analysis data on the basis of the score to generate a ranking data, and to output the ranking data,
wherein the related term is extracted from the words included in the plurality of pieces of first reference text analysis data, on the basis of a similarity degree or a proximity of distance between distributed representation vectors of the words and a distributed representation vector of the keyword,
wherein the weight of the keyword is a value based on an inverse document frequency of the keyword in the plurality of pieces of first reference text analysis data or the plurality of pieces of second reference text analysis data,
wherein the weight of the related term is a product of the weight of the keyword by 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, and
wherein the processing portion is configured to give a compiled weight to each of the keyword and the related term, to give a score to each of the plurality of pieces of second reference text analysis data on the basis of the compiled weight.

2. The document search system according to claim 1 ,
wherein the score is given to the second reference text analysis data including a word matching the keyword or the related term.

3. The document search system according to claim 1 ,
wherein the plurality of pieces of first reference text analysis data is the same as the plurality of pieces of second reference text analysis data.

4. The document search system according to claim 1 ,
wherein the related term is extracted using a distributed representation vector obtained by machine learning performed on distributed representation of the words included in the plurality of pieces of first reference text analysis data.

5. The document search system according to claim 1 ,
wherein each of the distributed representation vectors of the words is a vector generated with use of a neural network.

6. The document search system according to claim 1 ,
wherein extraction of the keyword included in text data comprises a morphological analysis on the text data to generate analysis data and extraction of the keyword from the analysis data, and
wherein the keyword is extracted from words included in the analysis data on the basis of an inverse document frequency in the plurality of pieces of first reference text analysis data or the plurality of pieces of second reference text analysis data.

7. The document search system according to claim 1 ,
wherein the weight is changeable by a user.

8. The document search system according to claim 1 ,
wherein the first reference text analysis data is data generated by performing morphological analysis on first reference text data, and
wherein the second reference text analysis data is data generated by performing morphological analysis on second reference text data.

9. The document search system according to claim 1 , further comprising:
an electronic device; and
a server,
wherein the electronic device comprises a first communication portion,
wherein the server comprises the processing portion and a second communication portion,
wherein the first communication portion is configured to supply the text data to the server through one or both of wire communication and wireless communication,
wherein the processing portion is configured to supply the ranking data to the second communication portion, and
wherein the second communication portion is configured to supply the ranking data to the electronic device through one or both of wire communication and wireless communication.

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

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

12. The document search system according to claim 1 ,
wherein the related term is configured to be included by synonyms of the keyword.

13. The document search system according to claim 1 ,
wherein the processing portion is configured to output the weight of the keyword and the weight of the related term.

14. A document search method comprising the steps of:
extracting a keyword included in text data;
extracting a related term of the keyword from words included in a plurality of pieces of first reference text analysis data;
giving a weight to each of the keyword and the related term;
giving a score to each of a plurality of pieces of second reference text analysis data on the basis of the weight;
ranking the plurality of pieces of second reference text analysis data on the basis of the score to generate a ranking data;
outputting the ranking data;
giving a compiled weight to each of the keyword and the related term; and
giving a score to each of the plurality of pieces of second reference text analysis data on the basis of the compiled weight,
wherein the related term is extracted from the words included in the plurality of pieces of first reference text analysis data, on the basis of a similarity degree or a proximity of distance between distributed representation vectors of the words and a distributed representation vector of the keyword,
wherein the weight of the keyword is a value based on an inverse document frequency of the keyword in the plurality of pieces of first reference text analysis data or the plurality of pieces of second reference text analysis data, and
wherein the weight of the related term is a product of the weight of the keyword by 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.

15. The document search method according to claim 14 ,
wherein the score is given to the second reference text analysis data including a word matching the keyword or the related term.

16. The document search method according to claim 14 ,
wherein the plurality of pieces of first reference text analysis data is the same as the plurality of pieces of second reference text analysis data.

17. The document search method according to claim 14 ,
wherein the related term is extracted using a distributed representation vector obtained by machine learning performed on distributed representation of the words included in the plurality of pieces of first reference text analysis data.

18. The document search method according to claim 14 ,
wherein each of the distributed representation vectors of the words is a vector generated with use of a neural network.

19. The document search method according to claim 14 ,
wherein the step of extracting the keyword included in the text data comprises the steps of:
performing morphological analysis on the text data to generate analysis data; and
extracting the keyword from wo

Assignees

• Semiconductor Energy Lab

Inventors

• Higashi Kazuki
• Momo Junpei

Classifications

• G06N3/09

  Supervised learning · CPC title

• G06N3/0499

  Feedforward networks · CPC title

• G06N3/08

  Learning methods · CPC title

• G06N3/063

  using electronic means · CPC title

• G06F40/279 Primary

  Recognition of textual entities · CPC title