UBER RIDES PREDICTION USING MACHINE LEARNING

Abstract

Accurately predicting ride demand is critical to improving Uber performance and increasing customer satisfaction. This paper investigates machine learning methods to predict Uber ride requests by analyzing historical data including time, location, and other factors such as weather. After preprocessing and feature engineering, various models (linear regression, decision trees, random forests, and GBR) are evaluated for their prediction performance. GBR emerged as the best model, achieving over 99% accuracy due to its ability to handle complex data and capture nonlinear relationships. This study highlights the importance of physical and environmental features in improving prediction accuracy, allowing Uber to allocate drivers more efficiently and reduce passenger waiting times. These findings provide insight into demand forecasting, improved management, and service quality of ridehailing platforms