Automatic acquisition for sensibility knowledge using co-occurrence relation
Tomoko Yoshinari, El-Sayed Atlam, Kazuhiro Morita, Kumiko Kiyoi, Jun‐ichi Aoe · International Journal of Computer Applications in Technology · 2008
When companies obtain customer opinions and requests from free-styled writings, sensibility expressions are important because they include personal subjectivity such as claims 'Too much dust'. Sensibility knowledge registered in the expressions is difficult to build manually. In order to reduce personal burden for building sensibility knowledge, this paper presents a method to acquire sensibility expressions automatically by using co-occurrence relation. According to experimental results of co-occurrence knowledge registered 1,300,000 terms, the rate of correct answers included in acquired sensibility expressions is 73.5%.