Journal of Big Data

Table 6 The compared results on the IRI dataset

From: ASENN: attention-based selective embedding neural networks for road distress prediction

Method	RMSE	\(R^2\)	MAE	MAPE
Random forest [5]	0.031	0.998	0.012	0.070
XGBoost [6]	0.025	0.999	0.008	0.038
LightGBM [7]	0.032	0.998	0.014	0.041
CatBoost [8]	0.041	0.998	0.014	0.036
MLP [3]	0.125	0.983	0.048	0.177
ResNet [14]	0.089	0.991	0.049	0.328
FT-transformer [10]	0.025	0.999	0.013	0.048
TabNet [9]	0.123	0.984	0.083	0.464
Ours	0.024	0.999	0.005	0.022

Back to article page