Impacts of Customer Feedback for Online-Offline Shopping using Machine Learning

Abstract

"Voice of the customer" is an extraordinary source for getting customer reviews and provides huge awareness into customer’s preferences towards a product or service. In an e-trade business, consumer reviews are very analytical. If a customer is purchasing a product through online sites, then the reviews of a product have an important role to play in decision making. The feedback of the products appears in the form of product ratings, consumer comments, and emotional reactions. Analyzing customer reviews (ACR) is important to evaluate customer satisfaction and would help the new customers in making better purchasing decisions. ACR also helps businesses to perform research, development, and focus on addressing customer issues after understanding customer’s opinions. The customer’s comments can be analyzed and classified as positive or negative and using classifiers to provide clear indications to the new customers regarding the products. For classification, the frequency of positive and negative words are calculated from the comments, the bag-of-word is formed based on customer reviews, and classification algorithms (Support Vector Machine, Logistic Regression, and Artificial Neural Network) are applied to classify comments as positive and negative reviews of customers. . This research work has collected the data related to Women’s Clothing E-commerce from the Kaggle dataset which includes 23486 reviews of the customer. This research is conducted to understand the customer perception towards online and offline shopping. This study come up with an analysis using a Machine Learning algorithm called SVM, LR, and ANN and simulated by using Python. Accuracy of SVM, LR, and ANN is also calculated where ANN has higher accuracy (i.e., 88%) than the SVM (i.e., 80%) and LR (i.e., 75%).