Document Level Sentiment Analysis Using Aspect Based Approach for Movie Reviews

Social media provides the best possible platform for sharing customer's view or comment about any product. The reviews provided by the customer's are generic in nature i.e., they give comment about the product as a whole. But, in some cases, the customer is concerned about a particular aspect of the product like camera quality, internal memory, size of the screen of the mobile etc., in such situation, the review provided for the whole product may not work correctly or useful. So, aspect based sentiment analysis (ABSA) approach is adopted in this work, where any particular aspect is considered and based on that the further analysis is carried out. The reviews are processed based on the aspect keyword and then, the processed reviews are then analyzed using machine learning technique. Support Vector Machine is used in this paper to classify the reviews in either positive or negative class. Various evaluation parameters are used to obtain the results, which inform the quality of the work.