Table 2.

The performance of the random forest algorithm from the testing phase using three different attribute/feature optimisation approaches. Values are in percentage.

Feature optimisation approach	Accuracy	Precision	Recall	F1-score	Error-rate
Random forest with features from UFS	77.50	66.67	94.12	78.05	22.50
Random forest with features from REF	72.50	63.64	82.35	71.19	27.50
Random forest with features from SFM	85.00	76.19	94.12	84.21	15.00