Table 7 Univariate ANOVAs
From: A framework to improve churn prediction performance in retail banking
Metric | Stage | df | Sum Sq | Mean Sq | F value | Pr(> F) |
|---|---|---|---|---|---|---|
PR-AUC | FE | 1 | 7.4743 | 7.4743 | 2678.21 | 0.0000 |
IDT-Over | 1 | 1.3519 | 1.3519 | 484.40 | 0.0000 | |
IDT-Under | 1 | 0.4983 | 0.4983 | 178.53 | 0.0000 | |
Classification Model | 1 | 4.6317 | 4.6317 | 1659.63 | 0.0000 | |
Residuals | 1595 | 4.4513 | 0.0028 | |||
Accuracy | FE | 1 | 10.3780 | 10.3780 | 4037.89 | 0.0000 |
IDT-Over | 1 | 0.9675 | 0.9675 | 376.43 | 0.0000 | |
IDT-Under | 1 | 0.7384 | 0.7384 | 287.30 | 0.0000 | |
Classification Model | 1 | 3.4610 | 3.4610 | 1346.59 | 0.0000 | |
Residuals | 1595 | 4.0994 | 0.0026 | |||
Recall | FE | 1 | 9.2224 | 9.2224 | 4260.11 | 0.0000 |
IDT-Over | 1 | 1.2777 | 1.2777 | 590.20 | 0.0000 | |
IDT-Under | 1 | 0.2430 | 0.2430 | 112.24 | 0.0000 | |
Classification Model | 1 | 4.4635 | 4.4635 | 2061.84 | 0.00000 | |
Residuals | 1595 | 3.4529 | 0.0022 | |||
Specificity | FE | 1 | 10.58 | 10.5839 | 4627.61 | 0.0000 |
IDT-Over | 1 | 0.8925 | 0.8925 | 390.21 | 0.0000 | |
IDT-Under | 1 | 0.7320 | 0.7320 | 320.05 | 0.0000 | |
Classification Model | 1 | 3.4350 | 3.4350 | 1501.89 | 0.0000 | |
Residuals | 1595 | 3.6480 | 0.0023 |