Email this article Information-Theoretic method for classification of texts
- Authors: Ryabko B.Y.1,2, Gus’kov A.E.1,3, Selivanova I.V.2,3
-
Affiliations:
- Institute of Computational Technologies
- Novosibirsk State University
- Russian National Public Library for Science and Technnology
- Issue: Vol 53, No 3 (2017)
- Pages: 294-304
- Section: Source Coding
- URL: https://journal-vniispk.ru/0032-9460/article/view/166433
- DOI: https://doi.org/10.1134/S0032946017030115
- ID: 166433
Cite item
Open Access
Access granted
Subscription Access
Abstract
We consider a method for automatic (i.e., unmanned) text classification based on methods of universal source coding (or “data compression”). We show that under certain restrictions the proposed method is consistent, i.e., the classification error tends to zero with increasing text lengths. As an example of practical use of the method we consider the classification problem for scientific texts (research papers, books, etc.). The proposed method is experimentally shown to be highly efficient.
About the authors
B. Ya. Ryabko
Institute of Computational Technologies; Novosibirsk State University
Author for correspondence.
Email: boris@ryabko.net
Russian Federation, Novosibirsk; Novosibirsk
A. E. Gus’kov
Institute of Computational Technologies; Russian National Public Library for Science and Technnology
Email: boris@ryabko.net
Russian Federation, Novosibirsk; Novosibirsk
I. V. Selivanova
Novosibirsk State University; Russian National Public Library for Science and Technnology
Email: boris@ryabko.net
Russian Federation, Novosibirsk; Novosibirsk
Supplementary files
