Sentiment information Extraction of comparative sentences based on CRF model

Wei Wang¹, Guodong Xin¹, Bailing Wang¹, Junheng Huang¹ and Yang Liu¹

School of Computer Science and Technology, Harbin Institute of Technology
150001, Harbin, China
{wwhit, gdxin,wbl, hithjh}@hit.edu.cn, lyylwhhit@126.com

Abstract

Comparative information mining is an important research topic in the sentiment analysis community. A comparative sentence expresses at least one similarity or difference relation between two objects. For example, the comparative sentence “The space of car A is bigger than that of car B and car C” expresses two comparative relations and . This paper introduces conditional random fields model to extract Chinese comparative information and focuses on the task of element extraction from comparative sentences. We use the conditional random fields model to combine diverse lexical, syntactic and semantic features derived from the texts. Experiments show that the proposed method is competitive and domain-independent, with promising results.

Key words

information extraction, comparative sentence, comparative element

Digital Object Identifier (DOI)

https://doi.org/10.2298/CSIS161229031W

Publication information

Volume 14, Issue 3 (September 2017)
Advances in Information Technology, Distributed and Model Driven Systems
Year of Publication: 2017
ISSN: 2406-1018 (Online)
Publisher: ComSIS Consortium

Full text

Download Available in PDF
Portable Document Format

How to cite

Wang, W., Xin, G., Wang, B., Huang, J., Liu, Y.: Sentiment information Extraction of comparative sentences based on CRF model. Computer Science and Information Systems, Vol. 14, No. 3, 823–837. (2017), https://doi.org/10.2298/CSIS161229031W