Turney, ACL 2001
From ScribbleWiki: Analysis of Social Media
Thumbs up or thumbs down? semantic orientation applied to unsupervised classification of reviews
An early paper and influential paper presenting an unsupervised approach to predicting the polarity of product reviews. The basic ideas are:
- Use patterns of POS tags to pick out phrases that are likely to be meaningful (e.g. ADJ NOUN might pick out "good service" or "delicious desserts").
- Use pointwise mutual information (PMI) to score the similarity of each phrase in a review with the two words "excellent" or "poor", and give an overall score polarity to each phrase based on the difference of its PMI with "excellent" to the PMI with "poor".
- Score the polarity of a review based on the total polarity of the phrases in it.
This approach was fairly successful on a range of review-classification tasks: it achieved accuracy of between 65% and 85% in predicting an author-assigned "recommended" flag for Epinions ratings for eight diverse "products", ranging from cars to movies. Many later writers used several key ideas from the paper, including: treating polarity prediction as a document-classification problem; classifying documents based on likely-to-be-informative phrases; and using unsupervised or semi-supervised learning methods.
An interesting follow-up paper is Turney and Littman, TOIS 2003 which focuses on evaluation of the technique of using PMI for predicting the semantic orientation of words.