The Efficacy of Predictive Methods in Financial Statement Fraud

Table 4

Descriptive statistics of the performance measures obtained from 1000 iterations segmented by different rate of fraudulent companies, and for training sample size of 300 and test sample size of 100.

		1% Fraud		3% Fraud		5% Fraud		7% Fraud		10% Fraud
		Prec.	recall	Prec.	recall	Prec.	recall	Prec.	recall	Prec.	recall
MFFNN	Avg.	0.955	0.949	0.955	0.947	0.957	0.946	0.957	0.945	0.957	0.940
	SD	0.033	0.035	0.034	0.036	0.034	0.036	0.033	0.036	0.033	0.038
	Min	0.805	0.796	0.805	0.813	0.805	0.800	0.829	0.816	0.829	0.796
	Max	1.000	1.000	1.000	1.000	1.000	1.000	1.000	1.000	1.000	1.000
PNN	Avg.	0.968	0.728	0.966	0.728	0.964	0.730	0.968	0.728	0.965	0.725
	SD	0.027	0.053	0.027	0.052	0.029	0.055	0.028	0.055	0.030	0.055
	Min	0.878	0.574	0.854	0.563	0.846	0.571	0.829	0.582	0.829	0.586
	Max	1.000	0.909	1.000	0.930	1.000	0.909	1.000	0.884	1.000	0.930
SVM	Avg.	0.892	0.898	0.888	0.898	0.884	0.897	0.885	0.900	0.885	0.897
	SD	0.049	0.045	0.050	0.045	0.052	0.046	0.052	0.047	0.050	0.048
	Min.	0.707	0.720	0.659	0.761	0.683	0.750	0.634	0.735	0.659	0.708
	Max	1.000	1.000	1.000	1.000	1.000	1.000	1.000	1.000	1.000	1.000
MLM	Avg.	0.874	0.875	0.877	0.872	0.876	0.874	0.875	0.877	0.878	0.873
	SD	0.055	0.049	0.055	0.048	0.055	0.049	0.056	0.049	0.054	0.050
	Min	0.683	0.720	0.610	0.694	0.683	0.667	0.634	0.708	0.659	0.673
	Max	1.000	1.000	1.000	1.000	1.000	1.000	1.000	1.000	1.000	1.000
DA	Avg.	0.897	0.854	0.895	0.861	0.891	0.869	0.890	0.878	0.891	0.886
	SD	0.050	0.049	0.051	0.048	0.052	0.048	0.052	0.048	0.051	0.048
	Min	0.683	0.685	0.683	0.698	0.659	0.727	0.707	0.694	0.659	0.732
	Max	1.000	1.000	1.000	1.000	1.000	1.000	1.000	1.000	1.000	1.000