Learning A Statistical Model Of Product Aspects For Sentiment Analysis

In this paper, we introduce a new methodology for modeling product aspects from a collection of free-text customer reviews. The proposal relies on a language modeling framework and is domain independent. It combines both a kernel based model of opinion words and a stochastic translation model between words to approach the aspect model of products. We also present a ranking-based methodology to model the sentiments expressed about the aspects. The experiments carried out over several collections of customer reviews show encouraging results in the modeling of product aspects and their sentiments even from individual customer reviews.