Product Price Analysis, Prediction and Product Recommendation: Insights from E-commerce Data using Machine Learning Techniques

The rapid expansion of online commerce presents significant opportunities for leveraging data to enhance user experience and drive business growth. This research aims to explore and apply various machine learning algorithms to improve predictive modelling and recommendation systems within E-commerce platforms. We emplotes a diverse dataset that includes product attributes, pricing details and customer feedback from major online platforms. Our approach involved using machine learning algorithms such as Generalized Linear Models (GLM), Lasso, Ridge regression, XGBoost, Decision Trees (DT) and Random Forest. We conducted exploratory analysis and forecast trends. Additionally, a recommendation system was developed to provide personalized product suggestions The Random Forest algorithm exhibited the highest classification accuracy of 94% and the lowest error rate of 6%, outperforming other models. Lasso and Ridge regression followed closely with accuracies of 93.8% and 93.7%, respectively, and error rates of 6.2% and 6.3%. The recommendation system effectively provided tailored suggestions, enhancing user satisfaction by aligning recommendations with individual preferences. Novelty: This work integrates advanced machine learning algorithms with a novel recommendation system, offering a comprehensive solution for optimizing product pricing and personalized recommendations in E-commerce. The approach not only improves predictive accuracy but also enhances customer engagement through tailored recommendations.