Research Article

Exploring an Interactive Value-Adding Data-Driven Model of Consumer Electronics Supply Chain Based on Least Squares Support Vector Machine

Table 3

Comparison of accuracies of four different models.

Actual values of test setPredictive values ()
Primary indicatorSecondary indicatorTertiary indicator
SVMBPELMLS-SVMSVMBPELMLS-SVMSVMBPELMLS-SVM

0.1020.101/0.0080.105/0.0290.089/0.1230.107/0.0530.110/0.0780.116/0.1410.119/0.1660.108/0.0590.107/0.0500.102/0.0000.1038/0.0180.109/0.071
0.0880.086/0.0260.108/0.2250.098/0.1110.084/0.0500.089/0.0130.088/0.0030.133/0.5110.085/0.0330.096/0.0850.109/0.2400.133/0.5110.088/0.001
0.0670.079/0.1790.081/0.2030.099/0.4800.066/0.0100.075/0.1150.054/0.1980.056/0.1660.072/0.0690.074/0.1090.059/0.1210.058/0.1340.067/0.001
Mean error0.0710.1520.2380.0380.0690.1140.2810.0540.0810.120.2210.024
Standard deviation0.0940.1070.2100.0240.0520.1000.2000.0190.0300.1200.2580.041

: is the predictive value; is the relative error.