Conversion of adverse data corpus to shrewd output using sampling metrics

Table 4 Classification results after oversampling

Classifier	Accuracy	Classes	Precision	Recall	F-Measure	Confusion matrix
Naïve bayes	87.89%	Low	0.852	0.907	0.879	a b < −- classified as 98 10 \| a = Low 17 98 \| b = High
		High	0.907	0.852	0.879
		Average	0.881	0.879	0.879
Multilayer perceptron	91.03%	Low	0.873	0.954	0.912	a b < −- classified as 103 5 \| a = Low 15,100 \| b = High
		High	0.952	0.870	0.909
		Average	0.914	0.910	0.910
SVM	88.79%	Low	0.849	0.935	0.890	a b < −- classified as 101 7 \| a = Low 18 97 \| b = High
		High	0.933	0.843	0.886
		Average	0.892	0.888	0.888
IBk	83.86%	Low	0.805	0.880	0.841	a b < −- classified as 95 13 \| a = Low 23 92 \| b = High
		High	0.876	0.800	0.836
		Average	0.842	0.839	0.838
Random forest	90.13%	Low	0.898	0.898	0.898	a b < −- classified as 97 11 \| a = Low 11,104 \| b = High
		High	0.904	0.904	0.904
		Average	0.901	0.901	0.901