A Hybrid Approach of Stepwise Regression, Logistic Regression, Support Vector Machine, and Decision Tree for Forecasting Fraudulent Financial Statements

Table 6

Summary of classification results.

Model	Type I error	Type II error	Hit ratio	Ranking
Logistic regression	9.52%	28.57%	80.95%	2
SVM	7.14%	48.81%	72.02%	3
DT C5.0	10.71%	17.85%	85.71%	1