Sentiment Mining Based on Products Reviews Using Machine Learning

Extracting and Analyzing Customer Sentiments using Sentiment Mining and Machine Learning on Product Reviews

Abstract

The increasing use of Internet and online activities (such as chatting, conferencing, surveillances, hotel and ticket reservation, B2C and B2B e-commerce, various social media platforms, blogging and micro-blogging, gets for us a very huge database of structured and unstructured data, referred to as Big Data, leading us to extract, transform, load, and analyse this data for interpretations. Such data can be examined using a mixture of Data Mining, Web Mining and Text Mining techniques in various real life applications. The base research by Rushleen et al [3] performed mining on tweets obtained on the Samsung Electronics twitter handle. The algorithm accurately analysis the positive, negative and moderate tweets. The algorithm accuracy is measured in terms of accuracy percentage and time complexity. These values are found to be 80 and O(mn) respectively. This paper focuses on extracting the features from product reviews taken from Cornell University Movies Review Database given by reviewers to state their opinions. This is done at aspect level of analysis using ontology. Then it determines whether they are positive or negative thereby giving a scaling system to identify the effectiveness of a product. The scaling system is in the form of a -5 to 0 to +5 marking system where negative and positive values indicate the customer sentiment. Output of such analysis is then summarized using machine learning techniques using Weka Tool.