Managing E-Reviews: A Performance Enhancement Technique Using Deep Learning

E-business has converted the entire world into a global market where client feedback plays a substantial role in improving the business's productivity as it helps the companies to meet the public demands. Reviews on the products' services helps in knowing the public requirements. Consumer reviews help the business owner to know the success of the product, consequently, the growth of the business as well as the future of the business. Various techniques had emerged in order to gather consumer responses. Sentiment analysis targets to garner the responses and categorize those reviews into some fixed patterns, for example- positive, negative or neutral. This methodology helps to study a large amount of data effectively, efficiently and quickly. This study aims to propose a system that accurately integrates text-based feedback into the thumbs up, thumbs down or non-partisan set. The proffered approach uses deep learning for sentimental analysis to figure out the text and convert it into valuable feedback. The proposed study incorporates a machine learning procedure called the TextEmbedNet Model, which incorporates word embedding, global average pooling, and Dense layers. The performance of the proposed work is measured using F1 Score, Precision, Recall and Specificity constraints. In order to produce the results dataset from Amazon has been taken into consideration.