Research Article

WSF2: A Novel Framework for Filtering Web Spam

Table 7

AUC results for C5.0 and SVM classifiers working together with regular expressions.

Class-imbalance ratio
1 : 171 : 81 : 41 : 21 : 1

C5.00.5620.6490.6510.6480.573
SVM0.5340.5900.6020.6040.624
C5.0 + SVM0.5790.6580.7130.6840.646
C5.0 + SVM + REGEX0.6730.7680.7980.7590.736