Analyzation of Sentiments of Product Reviews using Natural Language Processing

In order to analyze opinion mining from e-commerce websites, this paper discusses sentence-level sentiment analysis. To save time, many consumers use e-commerce websites like Amazon and eBay to purchase millions of things online. The product may already have a lot of reviews or comments from other customers who have already purchased it. In this paper, a Bayesian-network based framework is presented that makes use of product reviews imported from e-commerce websites for a specific product. The shipped product reviews are then automatically analyzed to produce untrained rules. After that, a Bayesian Network is constructed utilizing a set of untrained rules, evidence set and random variables. The newly generated reviews from e-commerce websites can subsequently be evaluated by the trained network. Additionally, the built-in model can be incrementally retrained to increase its resilience. Last but not the least, the framework is easily scalable to opinion mining from social networking sites on a variety of topics, and the model standards that are produced by this process are developed instantly and automatically.