The Hybrid Method for Accurate Patent Classification
- Authors: Yadrintsev V.V.1,2, Sochenkov I.V.1,3
-
Affiliations:
- Federal Research Center Computer Science and Control of the Russian Academy of Sciences
- Peoples’ Friendship University of Russia (RUDN University)
- Lomonosov Moscow State University
- Issue: Vol 40, No 11 (2019)
- Pages: 1873-1880
- Section: Article
- URL: https://journals.rcsi.science/1995-0802/article/view/206101
- DOI: https://doi.org/10.1134/S1995080219110325
- ID: 206101
Cite item
Abstract
This article is dedicated to stacking of two approaches of patent classification. First is based on linguistically-supported k-nearest neighbors algorithm using the method of search for topically similar documents based on a comparison of vectors of lexical descriptors. Second is the word embeddings based fastText, where the sentence (or a document) vector is obtained by averaging the n-gram embeddings, and then a multinomial logistic regression exploits these vectors as features. We show in Russian and English datasets that stacking classifier shows better results compared to single classifiers.
Keywords
About the authors
V. V. Yadrintsev
Federal Research Center Computer Science and Control of the Russian Academy of Sciences; Peoples’ Friendship University of Russia (RUDN University)
Author for correspondence.
Email: vvyadrincev@gmail.com
Russian Federation, Moscow, 119333; Moscow, 117198
I. V. Sochenkov
Federal Research Center Computer Science and Control of the Russian Academy of Sciences; Lomonosov Moscow State University
Author for correspondence.
Email: sochenkov@isa.ru
Russian Federation, Moscow, 119333; Moscow, 119991