Estimating Parking Availability using Decision Tree Regressor

Recent growth in the number of personal vehicles has aggravated the parking issue. Thus smart parking management system is a must in each metropolitan area. We use Decision Tree Regressor (DT) model to forecast the number of available parking places in the upcoming time. We tune the hyper-parameters max_depth of the DT on three datasets, CNR-A, CNR-B, and KLCC. Experiments show that our proposed model can be used for both small and large datasets. Our model outperforms the Random Forest Regressor (RFR), proposed in recent work, in term of MSE, MAE, RMSE, and R² metrics.