Demand Analysis of Taxi Passenger-carrying Hot Spot Areas Based on XGBoost Algorithm

In order to study the demand prediction of taxi passenger-carrying hot spot areas, firstly, the density-based DBSCAN clustering algorithm is used to identify the hot spot areas and delineate the boundaries of the hot spot areas. And combined with the demand characteristics of taxi, the demand prediction of hot spot areas is carried out by using XGBoost algorithm without Fourier eigenvalues and XGBoost algorithm with Fourier eigenvalues. By comparing the mean absolute percentage error (MAPE), the prediction accuracy of the latter is significantly improved.