Comparative Study of Machine Learning Approaches for Amazon Reviews
Abhilasha Singh Rathor, Amit Agarwal, Preeti Dimri · Procedia Computer Science · 2018
Sentiment analysis is a broadly employed method for finding and extracting the appropriate polarity of text sources using Natural language Processing (NLP) methods. This paper focuses on examining the efficiency of three machine learning techniques (Support Vector Machines (SVM), Naive Bayes (NB) and Maximum Entropy (ME)) for classification of online reviews using a web model using supervised learning methods. The reviews are divided as positive, neutral and negative. This is not only helpful for consumers those want to search the reviews of products prior to purchase but also for companies those want to observe the public’s reaction to their products. We have extracted Amazon Reviews using Amazon API. We have also used unigrams and weighted unigrams to train machine learning classifiers. The results have shown that machine learning algorithms work well on weighted unigrams and SVM has resulted maximum accuracy.