Cell Traffic Prediction Based on Convolutional Neural Network for Software-Defined Ultra-Dense Visible Light Communication Networks

Table 1

Results of forecasting performance

Model		2D model	HA	ARIMA	SVG
RMSE	Call in	9.72	44.00	12.92	20.57
	Call out	10.95	45.43	13.67	19.37
	SMS in	17.58	51.56	21.77	26.35
	SMS out	8.40	19.93	10.39	11.08
MAE	Call in	6.68	38.86	9.10	13.65
	Call out	7.68	40.55	9.76	13.57
	SMS in	12.33	45.18	14.87	16.50
	SMS out	6.14	16.87	7.41	7.59
MAPE	Call in	43.06	649.54	116.43	59.72
	Call out	47.24	713.40	83.04	57.52
	SMS in	64.02	455.01	71.13	83.31
	SMS out	68.32	421.39	86.37	86.42
DirAcc	Call in	0.83	0.62	0.71	0.67
	Call out	0.78	0.67	0.69	0.71
	SMS in	0.64	0.57	0.63	0.62
	SMS out	0.60	0.57	0.62	0.56