Extracting and Grounding Contextualized Sentiment Lexicons

A context-aware approach based on machine learning and lexical analysis identifies ambiguous terms and stores them in contextualized sentiment lexicons, which ground the terms to concepts corresponding to their polarity.

Abstract

Web intelligence applications track online sources with economic relevance such as customer reviews, news articles and social media postings. Automated sentiment analysis based on lexical methods or machine learning identifies the polarity of opinions expressed in these sources to assess how stakeholders perceive a topic. This paper introduces a hybrid approach that combines the throughput of lexical analysis with the flexibility of machine learning to resolve ambiguity and consider the context of sentiment terms. The context-aware method identifies ambiguous terms that vary in polarity depending on the context and stores them in contextualized sentiment lexicons. In conjunction with semantic knowledge bases, these lexicons help ground ambiguous sentiment terms to concepts that correspond to their polarity. This grounding paves the way for interlinking, extending, or even replacing contextualized sentiment lexicons with semantic knowledge bases. An extensive evaluation applies the method to user reviews across three domains (movies, products and hotels).