This project applies different classification techniques of sentiment analysis on reviews of yelp dataset to classify reviews as positive and negative.

Prashant Chhabra

Nathan Mots

Melina Sparks

In this project, we applied different classification techniques of sentiment analysis on reviews of yelp dataset to classify a reviews as positive and negative. We learned to use different algorithms for sentiment analysis and also played around with different settings preprocessing settings. We found that Support Vector Machine was slow but gave good accuracy and Multinomial Naive bayes gave good balance of speed and accuracy. Java Lingpipe library also gave us good accuracy and did not show a drastic decrease in accuracy on reducing number of test instances.

Please refer to presentation and report for details about analysis.