We consider the problem of classifying documents not by topic, but by overall sentiment, e.g., determining whether a review is positive or negative. Using movie reviews as data, we find that standard machine learning techniques definitively outperform human-produced baselines. However, the three machine learning methods we employed (Naive Bayes, maximum entropy classification, and support vector machines) do not perform as well on sentiment classification as on traditional topic-based categorization. We conclude by examining factors that make the sentiment classification problem more challenging. 1

This paper reports on the compilation of a large Persian Web corpus and the cyclic supervised development of a lexicon and lemmatizer. We discuss the strategies adopted in compiling the corpus as well as some of the challenges in processing and tokenizing it. We also present the word patterns developed for the lemmatizer and the algorithms designed for the supervised lexical acquisition. 1