Enhancing Domain Specific Sentiment Lexicon for Issue Identification

Abstract : The research work reported here is part of a larger project aimed at acquiring knowledge about issues in assembly, from documents. In order to do so, the first step is to identify the presence of issues. For this, sentiment analysis is proposed as a means. The presence of issues is proposed to be found by detecting negative sentiment. However, general English sentiment lexicons are not enough to detect negative sentiment in specialized domains, in this case, aircraft assembly. This paper studies an existing sentiment analysis tool, and creates an enhanced sentiment lexicon on the basis of the performance of the tool. The enhanced domain lexicon is proposed as the solution to identifying domain specific issues.