An Estimation Method of Descriptions of Estimation Condition in Travelers’ Reviews using Conditional Random Fields and Distributed Representation

Abstract

In sentiment analysis of travelers’ reviews, it is important to understand under what conditions the opinions were written by reviewers. We studied an estimation method of descriptions of estimation condition in travelers’ reviews using conditional random fields. In the previous research, 8 features in textual data were selected for conditional random fields. In this research, we proposed 2 new features and verified effectiveness of the proposed features by experiments. As the result, the F-measure was increased from 0.798 to 0.829.