Semantic Analysis and Implicit Target Extraction of Comments from E-Commerce Websites

In e-commerce websites, customers usually make comments, which include the properties of the product, the attitude to the vendor, express delivery information after buying the products. The information provides an important reference when others buy products in the website. In sentiment analysis, a finer-grained opinion mining approach focuses on not only the product itself as a whole but also product features, which can be a part or attribute of the product. Previous related research focuses on the explicit target mining but neglects the implicit ones. Whereas, the implicit features, which are implied by some words or phrases, are very significant and meaningful to express users´ opinion. In this paper, we propose a new approach to uniformly extract explicit and implicit opinion features from the comments. Our experimental results prove the effectiveness of the approach. The recall is improved, suggesting that taking the implicit features into consideration can extract more useful information, meanwhile, the precision is stable, not decreases.