Multi-layer perceptron based transfer passenger flow prediction in Istanbul transportation system

Estimating passenger movement in transportation networks is a critical aspect of public systems. It allows for greater understanding traffic patterns, as well efficient system evaluation and monitoring. could also help with precise timing to emergencies or important events, the improvement urban transport weaknesses service quality. The number transfer passengers demand Istanbul, Turkey's biggest most developed metropolis, was used construct real-world forecasting model this study. has been forecasted using popular machine learning methods such kNN (k-Nearest Neighbours), LR (Linear Regression), RF (Random Forest), SVM (Support Vector Machine), XGBoost MLP. dataset utilized made up hourly counts gathered at two stations Istanbul January 2020. Using MSE, RMSE, MAE R2 parameters, each model's experimental data have thoroughly evaluated. MLP more successfully other algorithms majority lines, according results.