A Word Vector Based Review Vector Method for Sentiment Analysis of Movie Reviews Exploring the Applicability of the Movie Reviews

摘要

Based on word embedding method, this paper presents a word vector based review vector method for sentiment analysis of movie reviews. As a result, it is achieved that 86.18% classification accuracy using the method. Meanwhile, the method is applicable to multiple languages such as Chinese and English, and it is extensible for larger scale contents as well. What's more, the influence of word vector dimensions on the sentiment analysis accuracy and the method's applicability on sentences of varied lengths are also discussed in this paper. The experimental result proved that the word vector based review method for sentiment analysis is not only an efficient and simple way to analyze emotional expression, but also has extensibility and applicability for comments in varied lengths and multiple languages.