Thumbs Up or Thumbs Down? Semantic Orientation Applied to Unsupervised Classification of Reviews

DMCA

Thumbs Up or Thumbs Down? Semantic Orientation Applied to Unsupervised Classification of Reviews (2002)

Cached

Download Links

Other Repositories/Bibliography

DBLP

by Peter Turney

Citations:

784 - 5 self

BibTeX

@INPROCEEDINGS{Turney02thumbsup,
    author = {Peter Turney},
    title = {Thumbs Up or Thumbs Down? Semantic Orientation Applied to Unsupervised Classification of Reviews},
    booktitle = {},
    year = {2002},
    pages = {417--424}
}

OpenURL

Abstract

This paper presents a simple unsupervised learning algorithm for classifying reviews as recommended (thumbs up) or not recommended (thumbs down). The classification of a review is predicted by the average semantic orientation of the phrases in the review that contain adjectives or adverbs. A phrase has a positive semantic orientation when it has good associations (e.g., "subtle nuances") and a negative semantic orientation when it has bad associations (e.g., "very cavalier"). In this paper, the semantic orientation of a phrase is calculated as the mutual information between the given phrase and the word "excellent" minus the mutual information between the given phrase and the word "poor". A review is classified as recommended if the average semantic orientation of its phrases is positive. The algorithm achieves an average accuracy of 74% when evaluated on 410 reviews from Epinions, sampled from four different domains (reviews of automobiles, banks, movies, and travel destinations). The accuracy ranges from 84% for automobile reviews to 66% for movie reviews.

Keyphrases

unsupervised classification semantic orientation applied mutual information average semantic orientation automobile review semantic orientation travel destination positive semantic orientation movie review word poor different domain word excellent subtle nuance average accuracy bad association simple unsupervised learning algorithm negative semantic orientation good association

DMCA