Full Text Patent Classification

Authors

Yadryntsev V. Sochenkov I.

Annotation

In this paper, the problem of patent documents classification is considered on the basis of a extended by nominal subgorups vector representation of full-text documents. The classification process begins by extracting keywords and phrases from the documents using by means of automatic text processing. Significant of keywords ans phrases are determining according to statistical measure. The topical similarity of documents based on vectors with keywords and phrases is estimating. In this work, the three lowest levels of international patent classification are used as a set of classes.

External links

Download the collection of the conference theses (PDF) from eLibrary (registration required): https://elibrary.ru/item.asp?id=35359532

Reference link

Ядринцев В. В., Соченков И. В. Full Text Patent Classification // Информационно-телекоммуникационные технологии и математическое моделирование высокотехнологичных систем: Материалы VIII Всероссийской конференции с международным участием (Москва, 16–20 апреля 2018). – М.: РУДН, 2018. С. 235–237.