Sentiment Analysis of Amazon Product Reviews using Deep Predictive Model

Reviews on Amazon are relevant in online shopping, as they inform customers about the quality, functionality, and performance of products, hence guiding customers to make purchasing decisions. To the seller, the reviews are really important in driving sales and improving visibility by Amazon's search algorithm. The paper presents sentiment analysis of Amazon reviews spanning sundry categories of products using Natural Language Processing and Deep Learning Techniques. The proposed model includes cognitive learning models and bag of words with TF-IDF approach to predict the review as either positive or negative. The outcome analysis has shown that Deep Neural Networks give the best training and testing accuracy of 94.6 % and 92.1% respectively in sentiment classification. The mean f-score obtained for various tokens was 91.8%. The research helps consumers to make better decisions and sellers to improve their product quality, making reviews more dependable and instructive resource for online shopping.