From online reviews, customers can obtain first-hand assessments of product information and direct supervision of their purchase actions. Meanwhile, manufacturers can obtain immediate feedback and opportunities to improve the quality of their products in a timely fashion. Thus, mining opinions from online reviews has become an increasingly urgent activity and has attracted a great deal of attention from researchers. To extract and analyze opinions from online reviews, it is unsatisfactory to merely obtain the overall sentiment about a product. In most cases, customers expect to find fine-grained sentiments about an aspect or feature of a product that is reviewed. To this end, this abstract proposes a novel approach based on the partially-supervised alignment model, which regards identifying opinion relations as an alignment process. Then, a graph-based co-ranking algorithm is exploited to estimate the confidence of each candidate. Finally, candidates with higher confidence are extracted as opinion targets or opinion words. Compared to previous methods based on the nearest-neighbor rules, our model captures opinion relations more precisely, especially for long-span relations. Compared to syntax-based methods, our word alignment model effectively alleviates the negative effects of par sing errors when dealing with informal online texts. In particular, compared to the traditional unsupervised alignment model, the proposed model obtains better precision because of the usage of partial supervision. In addition, when estimating candidate confidence, we penalize higher-degree vertices in our graph-based co-ranking algorithm to decrease the probability of error generation. Our experimental results on three corpora with different sizes and languages show that our approach effectively outperforms state-of-the-art methods.
You are here: / / The Word Alignment Model For Co-Extracting Opinion Targets And Opinion Words From Online Reviews