Abstract
As sentence level sentiment analysis having been studied extensively, it has been proven that the syntactic structure of a sentence usually holds important information for sentiment analysis, especially for handling polarity reversal. However, the previous attempts of adopting such structural information mainly focus on making certain predefined rules which requires large linguistic expertise of the rule-maker,and the procedure itself is often manually labored and time consuming. To solve this problem, in this paper we propose a novel simple vector model to represent a sentence’s syntactic structure and its prior sentiment information uniformly and rapidly. Experiment results show that our proposed approach performs well in COAE 2013 dataset, and could also be used for machine learning algorithms to extract more distinguish features automatically.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call