Subjectivity Detection in English and Bengali: A CRF-based Approach
Amitava Das, Sivaji Bandyopadhyay, Jadavpur Univers · 2009
With the proliferation of online reviews and sentiments the Web is becoming more and more useful and important information re-source for people. As a result, automatic opi-nion/sentiment mining has become a hot research topic recently. Extracting opinions from text is a hard semantic problem. Subjec-tivity Detection is studied as a text classifica-tion problem that classifies texts as either subjective or objective. This paper illustrates a Conditional Random Field (CRF) based Sub-jectivity Detection approach tested on English and Bengali multiple domain corpus to estab-lish its effectiveness over multiple domain perspective. The motivation is to develop ge-neric domain independent solution architec-ture for a less computerized language like Bengali. A relatively simple and less human interactive technique has been proposed for developing opinion mining resources for Ben-gali. The features used in the CRF-based clas-sifier could be extracted for any new language with minimum linguistics knowledge. The fi-nal classifier has resulted precision values of 76.08 % and 79.90 % for English and 72.16% and 74.6 % for Bengali for the news and blog domains respectively. 1