An Improved Feature Weighting Strategy in Chinese Text Categorization

DOI

10.2991/icmse-15.2015.40How to use a DOI?

Keywords

Feature Weighting, Inverse Category Frequency, TFIDF-ICF, Text Categorization

Abstract

In the process of document formalization, feature weight algorithm plays an essential role. It greatly interferes with the performance of the classifier. To improve the classic TF-IDF algorithm on its shortcomings that ignore feature distribution among the classes, we develop a new strategy to weight feature based on ICF and traditional TF-IDF. We have conducted a series of experiments on two text corpuses, namely the TanCorpV1.0 and Sogou, to analyze the performance of our strategy, which are described in the paper. Experimental results demonstrate the proposed strategy can to some extent improve the performance of text categorization.

Copyright

© 2015, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Jia Song
AU  - Sijun Qin
AU  - Pengzhou Zhang
PY  - 2015/12
DA  - 2015/12
TI  - An Improved Feature Weighting Strategy in Chinese Text Categorization
BT  - Proceedings of the 2015 6th International Conference on Manufacturing Science and Engineering
PB  - Atlantis Press
SP  - 202
EP  - 208
SN  - 2352-5401
UR  - https://doi.org/10.2991/icmse-15.2015.40
DO  - 10.2991/icmse-15.2015.40
ID  - Song2015/12
ER  -