Optimizing E-Commerce Supply Chains with Cat- Boosting: A Predictive Modelling Frame Work

Managing supply chains has become increasingly complex in today’s business environment. Traditional forecasting methods often fall short in providing accurate insights. To address this, Artificial Intelligence (AI) algorithms like CatBoost and Random Forest offer effective solutions. CatBoost, an ensemble-based machine learning model, handles categorical variables efficiently without requiring explicit encoding techniques, unlike other models. This study proposes a predictive modeling framework using CatBoost, along with Random Forest, to optimize supply chain processes. The models are tested on an e-commerce dataset, focusing on predicting sales, profit, and delivery times. The analysis shows that increasing the learning rate improves model performance, with CatBoost proving particularly effective for supply chain predictions.