Default or profit scoring credit systems? Evidence from European and US peer-to-peer lending markets

Financial Innovation

Table 3 Loan default classification comparison of credit and profit scoring models—Bondora

Model	Accuracy (%)	Specificity (%)	Sensitivity (%)
Panel A: Credit scoring models
LR	71.2	73.5	70.5
\(LR^{{\lambda_{\min } ,\alpha = 1}}\)	71.4	73.5	70.7
\(LR^{{\lambda_{\min } ,\alpha = 0}}\)	70.9	73.7	70.0
\(LR^{{\lambda_{\min } ,\alpha_{\min } }}\)	71.0	73.0	70.3
RFC	78.8	56.2	86.1
NNC	67.9	65.0	68.8
Panel B: Profit scoring models
LM	76.4	51.4	84.5
\(LM^{{\lambda_{\min } ,\alpha = 1}}\)	76.9	51.2	85.2
\(LM^{{\lambda_{\min } ,\alpha = 0}}\)	77.1	51.9	85.3
\(LM^{{\lambda_{\min } ,\alpha_{\min } }}\)	76.8	51.2	85.1
RFR	77.4	54.1	84.9
NNR	75.5	57.6	81.3

\(LR^{{\lambda_{\min } ,\alpha = 1}}\), \(LR^{{\lambda_{\min } ,\alpha = 0}}\), \(LR^{{\lambda_{\min } ,\alpha_{\min } }}\) are lasso, ridge and elastic net version of logistic regression, \(LM^{{\lambda_{\min } ,\alpha = 1}}\), \(LM^{{\lambda_{\min } ,\alpha = 0}}\), \(LM^{{\lambda_{\min } ,\alpha_{\min } }}\) are lasso, ridge and elastic net version of linear regression
RFC random forest classification, NNC neural network classification, RRR random forest regression, NNR neural network regression