Strategic Price Optimization: Machine Learning's Role in E-Commerce Dynamics

In the context of intense competition in the digital age, e-commerce companies must be able to adapt their pricing algorithms in real time in order to remain competitive. This research proposes an advanced dynamic pricing model that employs machine learning to effectively vary price levels. Through exploratory data analysis (EDA) on a large dataset, we identify correlations between pricing patterns and consumer responses. To this end, we incorporate Internet of Behaviors (IoB) technology to measure consumer price sensitivity. The study employs feature engineering and several regression algorithms to identify the most effective models. The results indicate that both Linear Regression and Random Forest Regression emerge as the top performing models. The Linear Regression model achieved an R-squared of 98.714% and a root mean square error (RMSE) of 7.955, while the Random Forest Regression exhibited a slightly higher R-squared of 98.7626% with an RMSE of 7.998. Other noteworthy models include Ridge Regression, which achieved an R-squared of 98.7135% and an RMSE of 7.960. The deployment of the model is facilitated by the use of a Streamlit interface, which provides the capability to generate robust, data-driven pricing recommendations in real time. This project builds upon existing literature by proposing an operational model for online businesses to make informed decisions based on comprehensive data analysis of user insights. The model demonstrates significant improvements over traditional analysis techniques.