. 2017 Nov 9;2017:8734214. doi: 10.1155/2017/8734214

Table 6.

The results of compare forecasting models under percentage spilt (dataset partition into 66% training data and 34% testing data) after variable selection.

	Methods	Index	RBF Network	Kstar	Random Forest	IBK	Random Tree
After variable selection	Delete the rows with missing data	CC	0.033	0.638^a	0.729^a	0.251	0.545^a
		MAE	0.182^a	0.121^a	0.111^a	0.199	0.135^a
		RMSE	0.229	0.176^a	0.156^a	0.287	0.212^a
		RAE	0.992^a	0.657^a	0.602^a	1.083	0.736^a
		RRSE	1.007	0.775^a	0.688^a	1.262	0.935^a
	Series mean	CC	0.107^a	0.661^a	0.739^a	0.242^a	0.551
		MAE	0.172^a	0.107^a	0.101^a	0.179^a	0.129
		RMSE	0.221^a	0.167^a	0.151^a	0.268^a	0.205
		RAE	0.988^a	0.615^a	0.579^a	1.027^a	0.740
		RRSE	0.995^a	0.753^a	0.678^a	1.208^a	0.923
	Linear	CC	0.105^a	0.666^a	0.735^a	0.258^a	0.596^a
		MAE	0.173^a	0.106^a	0.100^a	0.175^a	0.120^a
		RMSE	0.221^a	0.166^a	0.151^a	0.266^a	0.196^a
		RAE	0.987^a	0.606^a	0.572^a	1.002^a	0.683^a
		RRSE	0.995^a	0.748^a	0.681^a	1.198^a	0.883^a
	Median of nearby points	CC	0.106^a	0.666^a	0.740^a	0.264^a	0.553
		MAE	0.173^a	0.107^a	0.100^a	0.177^a	0.127
		RMSE	0.221^a	0.166^a	0.151^a	0.266^a	0.207
		RAE	0.987^a	0.611^a	0.571^a	1.013^a	0.723
		RRSE	0.995^a	0.747^a	0.677^a	1.195^a	0.932
	Mean of nearby points	CC	0.1059^a	0.667^a	0.745 ^ab	0.249^a	0.540^a
		MAE	0.173^a	0.107^a	0.099 ^ab	0.179^a	0.129^a
		RMSE	0.221^a	0.166^a	0.149 ^ab	0.268^a	0.214^a
		RAE	0.987^a	0.611^a	0.565 ^ab	1.025^a	0.735^a
		RRSE	0.995^a	0.747^a	0.672 ^ab	1.207^a	0.962^a
	Regression	CC	0.107^a	0.663^a	0.739^a	0.242^a	0.559^a
		MAE	0.172^a	0.106^a	0.101^a	0.179^a	0.126^a
		RMSE	0.221^a	0.167^a	0.151^a	0.268^a	0.200^a
		RAE	0.987^a	0.610^a	0.581^a	1.027^a	0.723^a
		RRSE	0.994^a	0.752^a	0.678^a	1.207^a	0.900^a

a denotes after variable selection with enhancing performance; b denotes the best performance among 5 models after variable selection.