Unsupervised extraction of popular product attributes from web sites

We develop an unsupervised learning framework for extracting popular product attributes from different Web product description pages. Unlike existing systems which do not differentiate the popularity of the attributes, we propose a framework which is able not only to detect concerned popular features of a product from a collection of customer reviews, but also to map these popular features to the related product attributes, and at the same time to extract these attributes from description pages. To tackle the technical challenges, we develop a discriminative graphical model based on hidden Conditional Random Fields. We have conducted experiments on several product domains. The empirical results show that our framework is effective.