Comparative Polarity Analysis on Amazon Product Reviews Using Existing Machine Learning Algorithms

Abstract

Opinion mining has emerged as an active domain among the research fraternity because an enormous amount of heterogeneous user data is continuously increasing every day via www, viz., e-commerce websites, social networks, discussion forums, blogs etc. Intentions are expressed in a different way with different vocabulary, short forms, and jargon making the data massive and disorganized. It has turned out an exciting new trend in social media with a scope of practical applications like analyzing marketing strategy, political analysis, social media analysis, financial analysis to take an effective decision based on user's feedback. With that being said, there are many unsolvable issues and challenges still existing that keeps the field more dynamic. As we know, it is very difficult to understand human language and more complex to program the machine to analyze the user's context. The proposed work aims at evaluating the sentiments of Amazon product reviews by using programming languages with the help of existing natural language processing APIs. Our aim is to categorize the sentiments of a Meta dataset, parse and protrude the comparative accuracy levels. Senti - A new algorithm is proposed to analyze the sentiments accurately which outperforms the existing APIs.