A Novel Hybrid Network Traffic Prediction Approach Based on Support Vector Machines

Table 1

The result of evaluation metrics of the four forecasting methods (inflow NTD).

Date	RMSE				MAE				MAPE
	BSVM	PSVM	ESVM	EPSVM	BSVM	PSVM	ESVM	EPSVM	BSVM	PSVM	ESVM	EPSVM
Monday	20.082	17.033	11.994	9.262	16.822	14.226	9.298	7.369	0.0893	0.0747	0.0495	0.0379
Tuesday	18.086	14.846	12.202	13.892	15.834	11.772	10.274	11.649	0.0828	0.0625	0.0543	0.0608
Wednesday	20.668	20.784	20.555	12.335	16.825	17.086	16.185	10.067	0.0797	0.0808	0.0766	0.0482
Thursday	23.445	16.836	16.226	11.427	19.741	12.806	12.776	9.878	0.0948	0.0615	0.0609	0.0481
Friday	16.729	21.573	15.683	13.456	13.962	17.455	11.897	10.549	0.0654	0.0825	0.0571	0.0487
Saturday	21.784	22.186	18.898	13.915	18.422	19.146	14.877	10.977	0.0882	0.0904	0.0704	0.0524
Sunday	22.104	14.935	19.864	15.537	18.288	12.056	17.359	12.459	0.0776	0.0517	0.0741	0.0530
Whole week	20.414	18.313	16.489	12.832	17.128	14.935	13.238	10.421	0.0825	0.0720	0.0633	0.0498