The Topic Tracking Based on Semantic Similarity of Sememe’s Lexical Chain

DOI

10.2991/sekeie-14.2014.28How to use a DOI?

Keywords

Topic tracking; Semantic Similarity;Vector Space Model; Lexical chain; Sememe

Abstract

In method of Semantic similarity calculating, the major is based on VSM(Vector Space Model).It has aroused significant research attention in recent years due to its advantage in topic tracking. In this paper a modified VSM, namely Semantic Vector Space Model, is put forward. To establish the model, numerous lexical chains based on HowNet are first built, then sememes of the lexical chains are extracted as characteristics of feature vectors. Afterwards, initial weight and structural weight of the characteristics are calculated to construct the Semantic Vector Space Model, encompassing both semantic and structural information. The initial weight is collected from word frequency, while the structure weight is obtained from a designed calculation method. Finally, the model is applied in web news topic tracking with satisfactory experimental results, conforming the method to be effective and desirable.

Copyright

© 2014, 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  - Jing Ma
AU  - Fei Wu
AU  - Chi Li
PY  - 2014/03
DA  - 2014/03
TI  - The Topic Tracking Based on Semantic Similarity of Sememe’s Lexical Chain
BT  - Proceedings of the 2nd International Conference on Software Engineering, Knowledge Engineering and Information Engineering (SEKEIE 2014)
PB  - Atlantis Press
SP  - 118
EP  - 121
SN  - 1951-6851
UR  - https://doi.org/10.2991/sekeie-14.2014.28
DO  - 10.2991/sekeie-14.2014.28
ID  - Ma2014/03
ER  -