Table 5.

Driving behavior identification for different methods when setting $T_x$ to 60 samples and $\Delta t$ to 6 samples.

Model	${\mathit{T}}_{\mathit{x}}=60, \mathsf{\Delta }\mathit{t}=6$
	Accuracy		AUC		F1 Score
	Mean	Std	Mean	Std	Mean	Std
KNN	0.9033	0	0.947	0	0.9045	0
DecisionTree	0.8543	0.0062	0.9213	0.004	0.8537	0.0069
RandomForest	0.8739	0.0044	0.9359	0.0025	0.8934	0.0034
DeepConvGRU-Attention	0.9836	0.0015	0.9978	0.001	0.9836	0.0015
DeepConvLSTM-Attention	0.9786	0.0068	0.9978	0.0006	0.9787	0.0068
DeepConvGRU	0.9772	0.0062	0.9968	0.0008	0.9772	0.0062
DeepConvLSTM	0.9519	0.0186	0.9944	0.0013	0.9497	0.019
CNN	0.9568	0.0072	0.9984	0.0002	0.9567	0.0073
LSTM-15	0.993	0.0015	0.9996	0.0001	0.9929	0.0015
DNN-45	0.9395	0.0358	0.9682	0.0281	0.9315	0.0493