“Dimension Reduction: Feature Subset” Method for Selecting the Best Index Combination in Reputation Evaluation of Crowdsourcing Participants

Table 7

Accuracy and dispersion degree of the selection classifier of the reputation evaluation index of crowdsourcing participants.

Selection classifier	Minimum value	Maximum value	Mean value	Standard deviation	Skewness coefficient	Kurtosis coefficient
ReliefF-DT	0.898	0.915	0.906	0.006	0.406	−0.581
ReliefF-SVM	0.906	0.922	0.915	0.005	−0.485	−0.553
ReliefF-BPNN	0.905	0.922	0.913	0.006	0.576	−0.282
ReliefF-RBFNN	0.886	0.911	0.899	0.008	−0.014	−1.040
ReliefF-KNN	0.882	0.911	0.896	0.009	0.031	−0.335
ReliefF-NB	0.879	0.905	0.889	0.009	0.544	−0.474

The best result of the classifier is shown in bold under different measures.