Paper Title
BIKE RENTAL COUNT PREDICTION BASED ON MACHINE LEARNING MODELS

Abstract
Abstract - In today's era, where climate change and sustainable development are major issues. Bike sharing system becomes one the important aspect of society. With the growing population and traffic, the bike rental system provides an efficient and cheap mode of transportation. The goal of this study is to foresee the demand for bike rentals by integrating historical usage patterns and weather data using three different regression models: (a) Linear Regression (b) Polynomial Regression (c) Gradient Boosting. Initially, the multiple linear regression model was developed through conventional techniques. However, upon assessing the performance of the model against actual values, it was discovered that its predictive accuracy was comparatively less accurate. The present study suggests the utilization of a Gradient Boosting Regressor and polynomial feature model to enhance the outcome. After comparing the performance of the three models, it was concluded that the Gradient Boosting Regression model exhibited the highest accuracy and the best RMSE value. Keywords - Bike Sharing System, Linear Regression, Gradient Boosting, Polynomial Feature, RMSE