Fast and Accurate Sentiment Classification Using an Enhanced Naive Bayes Model

摘要

We have explored different methods of improving the accuracy of a Naive Bayesclassifier for sentiment analysis. We observed that a combination of methodslike negation handling, word n-grams and feature selection by mutualinformation results in a significant improvement in accuracy. This implies thata highly accurate and fast sentiment classifier can be built using a simpleNaive Bayes model that has linear training and testing time complexities. Weachieved an accuracy of 88.80% on the popular IMDB movie reviews dataset.