Biographies, Bollywood, Boom-boxes and Blenders: Domain Adaptation for Sentiment Classification

TLDR

Automatic sentiment classification has been extensively studied and applied, yet sentiment expression varies across domains, making domain‑specific annotation impractical. The study investigates domain adaptation for sentiment classifiers by focusing on online reviews of different product types. The authors extend the structural correspondence learning (SCL) algorithm to sentiment classification, applying it to online reviews across product domains. The adapted SCL reduces relative error by an average of 30% over the original SCL and 46% over a supervised baseline, and a domain similarity measure correlates with adaptation potential, enabling selection of domains whose classifiers transfer well to many others.

Abstract

Automatic sentiment classification has been extensively studied and applied in recent years. However, sentiment is expressed differently in different domains, and annotating corpora for every possible domain of interest is impractical. We investigate domain adaptation for sentiment classifiers, focusing on online reviews for different types of products. First, we extend to sentiment classification the recently-proposed structural correspondence learning (SCL) algorithm, reducing the relative error due to adaptation between domains by an average of 30% over the original SCL algorithm and 46% over a supervised baseline. Second, we identify a measure of domain similarity that correlates well with the potential for adaptation of a classifier from one domain to another. This measure could for instance be used to select a small set of domains to annotate whose trained classifiers would transfer well to many other domains.