Return To Origin (RTO) is a major issue faced by e-commerce businesses. We develop a prototype to predict the product return made by the customers. Our model predicts the return of the products in advance so that necessary measures can be carried out to minimize the overall operational and financial losses. The research focuses on obtaining a normalized return score using three entities: consumer’s return behavior, product return rate from Vendor’s, and the feedback submitted by the customers. A hypothesis is set with the conditional statements and the obtained return score is checked to the range in which it falls and conclusions are finally driven out. The part of the system is implemented in python using a real-time dataset. The overall intention of this research is to decrease the return rate. The increase in RTO affects the revenue of the e-commerce firm and also for the vendors, so this model can assist them in decision-making for minimizing their losses.

Consumer Behavior, Sentiment Analysis, Natural Language Processing, Polarity, Predictive Model, E-commerce