Sentiment Analysis on Womens E-Commerce clothing
Welcome! In this project we implemented sentiment analysis techniques to analyze over 23 thousand E-Commerce reviews.
The first approach used was the Lexicographic Approach. We used a listing of negative and a listing of positive words as a basis for the analysis. With the assistance of Python, we counted the number of words from each review that fell within the positive or negative listings and compared the results.
The second approach used was the supervised machine learning approach. The data was split into two categories : training and testing. The training data was vectorized as a set of all words and the features were provided to a NaiveBayes Classifier.
The results of the two approaches were reviewed and compared.