Logistic Regression-Based Classification for Reviews Analysis on E-Commerce Based Applications

摘要

In recent years, many web applications play an important role that too online-commerce-based applications are ruling the web-based application by attracting the customer based on their products and providing more discount on the product. Customers buy the products online because of low-cost and the quick delivery of the products, and therefore for measuring the quality of the product, the user's comments on the different products play an important role which can be used for getting useful information on the product. In this paper, we do extract the data set from various E-commerce applications then by applying various methods like data pre-processing, classification and clustering on the products API we can able to analyse the reviews of the customers on different products. In this paper, we have created the product API for 100 instances of products by considering different attributes like product id, product name, total number of comments, rating and the text analysis. In this paper, first we apply the data pre-processing to clean the data set and then apply the logistic regression-based classification on the attributes taken, then form a cluster based on positive and negative reviews provided by the users and determining the accuracy using logistic regression technique and then comparing with decision tree technique accuracy with proposed logistic regression technique we achieve an improved accuracy.