Opinion mining based on feature-level

An important task of opinion mining is to extract people's opinions on features of an entity. However, for the same feature, people can express it with many different words or phrases. To produce a useful summary, these words and phrases, which are domain synonyms, need to be grouped into the same feature group. Moreover, the sentiment relatedness between the features and opinions is usually complicated. For many cases, product feature words are implied by the opinion words in reviews. A novel method is proposed to deal with the feature-level opinion mining problems. More Specially, 1) the proposed method considers the explicit features and the implicit features. 2) the opinion words are divided into two categories, vague opinion words and clear opinion words, to identify the implicit features and cluster the features. The feature clustering depends on three aspects: the corresponding opinion words, the similarity of the features and the structures of the features. Moreover, the context information is used to enhance the clustering in the procedure, which is proved to be useful in clustering. The experimental results demonstrate that our method performs well.