Prediction of seasonal bike rental counts using a GBM model optimized with bat algorithm

Abstract

To ensure effective resource allocation for urban bike demand, it is crucial to accurately predict shared bike rental counts. This prediction process was carried out using the Gradient Boosted Machine (GBM) method optimized with the Bat Algorithm (BA). To demonstrate the effectiveness of the proposed model, its performance was compared with different methods such as Decision Tree (DT), k -Nearest Neighbors (KNN), and Multi -Layer Perceptron (MLP). For this comparison, metrics such as Mean Absolute Error (MAE), Mean Squared Error (MSE), Root Mean Squared Error (RMSE), and R -squared (R 2 ) were employed. The best results were achieved by BA-GBM with values of 1.8665 MAE, 2.9588 MSE, 8.7545 RMSE, and 0.9264 R 2 . Additionally, the features with the most and least impact on bike rental prediction were identified. The most influential features were found to be temperature and time of day, while the least influential features were snowfall and year.