Empowering E-Commerce Decisions Through AI-Enhanced Review Analytics

Due to huge customer market product reviews are large in number and complicated. Thus, understanding the actual essence of the feedback is difficult. The system proposed by us is complex but it tackles the issues faced by previous systems. The paper serves as guideline for creating such system. For gathering data the method of web scraping is used, followed by preprocessing data using Natural Language Processing (NLP) techniques. For prediction purpose deep learning models LSTM and Convolutional Neural Networks (CNNs) are used along with Classification model XGBoost. Through our system we have presented the results visually so that consumers as well as Business stakeholders can take actionable insights. The system has great potential in variety of sectors as it can capture the sentiment of the reviews for a product and also its genuineness with precise accuracy. It has achieved fabulous training accuracy of 89.19% and validation accuracy of 81.49%. It represents the true sentiment of the customer feedback eliminating the fake reviews.