Fig. 1.

Performance of different classifiers and the GBDT method with different number of boosting stages. The left vertical coordinate is for average accuracy, while the horizontal coordinate is for classifier: KNN, SVM, NN, GBDT (with N = 50), GBDT (with N = 100), GBDT (with N = 200), GBDT (with N = 500), GBDT (with N = 1000) and GBDT (with N = 2000), respectively. Where L = 0.1, M = 3, F = 1. GB50, GB100, GB200, GB500, GB1000 and GB2000 represent the GBDT ensemble classifier with 50, 100, 200, 500, 1000 and 2000 boosting stages, respectively