RUS  ENG
Full version
JOURNALS // Informatsionnye Tekhnologii i Vychslitel'nye Sistemy // Archive

Informatsionnye Tekhnologii i Vychslitel'nye Sistemy, 2013 Issue 1, Pages 33–42 (Mi itvs105)

DATA PROCESSING

A method for topic clustering for large science publication collections

D. A. Deviatkinab, R. E. Suvorovab, I. V. Sochenkovba

a LLC "TSA"
b Institute for Systems Analysis of Russian Academy of Sciences

Abstract: The article covers research in the field of topic clustering for large science publication collections. Demands of developing such methods are considered. The method and the algorithm for topic clustering for large science publication amounts are presented. A comparison of the proposed method with classic clustering approaches is performed.

Keywords: text clustering, text classification, lexical descriptors, text spectral index, inverted spectral index, TF, IDF, topic importance characteristic, assessment of clustering methods.



© Steklov Math. Inst. of RAS, 2024