Predicting Consumer Behaviour from E-Commerce Platform Product Reviews: A Machine Learning Approach to Sentiment Analysis

Abstract

Given the recent advancements in web technology, users make use of the online services and offer their feedback thereby creating a vast amount of data available on the internet. To enable improved decision-making, it is essential to study, analyze, and organize this vast amount of user opinion, perspectives, feedback, and suggestions available through online resources. Nevertheless, the majority of e-commerce websites encounter challenges in comprehensively analyzing each individual review owing to the sheer volume of user feedback they receive. The aim of this study is to conduct sentiment analysis of product review from e-commerce platforms using machine learning to predict consumer behaviour so as to make informed decision. The methodology used was a sentiment analysis using models developed in Python with machine learning algorithms (Support Vector Machine and Naive Bayes) in order to compare the performance accuracy. The model was trained and tested on a dataset, and the performance of the two supervised machine learning algorithms for sentiment classification was compared. Accordingly, the results of the experiment showed that the Support Vector Machine technique outperformed Naive Bayes, obtaining accuracy levels of 97%. The study also demonstrates the significance of carrying out further research by exploring different machine-learning techniques and evaluating their performance on local market datasets. Additionally, it was observed that utilizing more data might improve the model's classification and prediction accuracy.