Department

Computer Science Department

Degree Name

BS in Computer Science

Date

12-2016

Advisor(s)

Franz Kurfess

Abstract/Summary

The goal of this project is to find a way to analyze reviews and determine the sentiment of a review. It uses various machine learning techniques in order to achieve its goals such as SVMs and Naive Bayes. Overall the purpose is to learn many different machine learning techniques, determine which ones would be useful for the project, then compare the results. Research is the foremost goal of the project, and it is able to determine the better algorithm for review classification, naive bayes or an SVM. In addition, an SVM which actually gave review’s scores rather than just classifying a review as negative or positive was created.

The next step is benchmarking both the SVM and Naive Bayes approach. Both algorithms were run through various tests of different input sizes and different reviews. The algorithms were timed for speed in prediction, classification, and most importantly accuracy. Overall, both approaches are great for different reasons that will be explored throughout the project.

Share

COinS