Extracting the Component Composition Data of Inventions from Russian Patents using Dependency Tree Analysis

Result code in IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00216208%3A11320%2F23%3AP85HG2W3" target="_blank" >RIV/00216208:11320/23:P85HG2W3 - isvavai.cz</a>

Result on the web

<a href="https://www.scopus.com/inward/record.uri?eid=2-s2.0-85162852443&doi=10.1109%2fICIEAM57311.2023.10139170&partnerID=40&md5=1ba465dae2c137c40b44d92621dc4334" target="_blank" >https://www.scopus.com/inward/record.uri?eid=2-s2.0-85162852443&doi=10.1109%2fICIEAM57311.2023.10139170&partnerID=40&md5=1ba465dae2c137c40b44d92621dc4334</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1109/icieam57311.2023.10139170" target="_blank" >10.1109/icieam57311.2023.10139170</a>

Result language

angličtina

Original language name

Extracting the Component Composition Data of Inventions from Russian Patents using Dependency Tree Analysis

Original language description

"The paper presents a methodology for extracting device components and relationships between them from the Russian-language patent claims. Information about the components of the device is the most useful and important part. It can be used in various tasks of patent analysis. The objective of this study is to evaluate the the quality of data extraction using dependency tree analysis for Russian language. The dependency tree for a sentence is the result of syntactic parsing by natural language processing tools. There are several parsers were chosen for comparison: UdPipe, Stanza, DeepPavlov, spaCy and Trankit. The output data are presented in the form of SAO structures (Subject-Action-Object). The quality of data extraction has been evaluated using precision, recall and F1 metrics. For this purpose, 20 patent claims with 252 SAO structures were manually marked. Under the current methodological constraints, on the test dataset, at best we are able to extract 81% of the SAO structures according to the recall metric with a non-strict data evaluation, i.e. without considering the completeness of a noun phrases. The F1-measure is lower and ranges from 48% to 69% depending on evaluation type. The current level of parsers efficiency in the investigated area is summarized. The results can be useful for developing efficient approaches to extracting structured data from patent arrays. © 2023 IEEE."

Czech name

—

Czech description

—

Type

D - Article in proceedings

CEP classification

—

OECD FORD branch

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Project

—

Continuities

—

Publication year

2023

Confidentiality

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Article name in the collection

"Proc. - Int. Conf. Ind. Eng., Appl. Manuf., ICIEAM"

ISBN

978-166547595-2

ISSN

—

e-ISSN

—

Number of pages

5

Pages from-to

1030-1034

Publisher name

Institute of Electrical and Electronics Engineers Inc.

Place of publication

—

Event location

Cham

Event date

Jan 1, 2023

Type of event by nationality

WRD - Celosvětová akce

UT code for WoS article

—