A neural network based approach for sentimental analysis on amazon product reviews

Abstract

A Sentiment is an opinion or thoughts stimulated by human feelings. On the other hand, the growth of internet technology allows everyone to share their opinions on social-media or micro-blogs. That’s how Sentiment Analysis has come into the picture in recent days. Mainly sentiment analysis contributes to the online products, political victory, film hits and celebrity dominations on social networks. This paper focalizes on product testimonials for online shopping websites. The online product reviews datasets are employed in this study which is taken from Amazon.com. From that dataset, the customer reviews have been analyzed through NLP and the corresponding sentiments have been scrutinized. Machine learning is a niche technology that is implemented in most applications nowadays. The sentiment analysis system is based on Machine Learning Prediction Analysis where Neural Networks are involved. Neural Networks is a powerful algorithm in machine learning techniques, which has got an architecture similar to the human brain system. MLP Neural Networks is a kind of Neural Network algorithm that is been used in this sentiment analysis. Neural Networks works on a larger dataset and sentiment analysis is more efficient compared to other machine learning algorithms such as the KNN algorithm, Naïve Bayes. Hence the Amazon products recommendation system is built with Neural Networks for effective performance