Attention based multi-component spatiotemporal cross-domain neural network model for wireless cellular network traffic prediction

EURASIP Journal on Advances in Signal Processing

Table 4 Performance comparison of various models and other models on the call dataset

Model	Multiple factors							RMSE	MAE	R²
Model	F₀	F_r	F_d	F_w	F_m	F_s	F_c	RMSE	MAE	R²
LR	√							50.8571	25.903	0.8102
SVR	√							45.8714	21.9142	0.8438
LSTM	√							44.4286	21.2136	0.8521
DenseNet	√							44.3864	19.144	0.8602
STMNet	√					√		39.7341	17.8913	0.8896
STNet	√				√	√		37.0857	16.8159	0.9129
STCNet	√				√	√	√	33.207	16.2393	0.9204
att-MCSTCNet (Conv-LSTM)	√	√	√	√	√	√	√	22.7051	14.2622	0.9628
att-MCSTCNet (Conv-GRU)	√	√	√	√	√	√	√	20.9656	13.1964	0.9683