Skip to main content

Table 3 Hypothesis contrast (p value) about difference in mean values

From: Measuring the model risk-adjusted performance of machine learning algorithms in credit default prediction

 

Statistics

Technology

Market conduct

Capital versus credit scoring

0.0181

0.1897

0.003

Capital versus provisions

0.0133

0.0180

0.5618

Credit scoring versus provisions

0.9404

0.1097

0.0128

  1. Bold means that differences between model uses are statistically significant at 10% of significance level