Financial Risk Manager Part 1

Explanation

Based on the confusion matrix provided:

• TN (True Negative) = 100: Correctly predicted no default when there was no default
• FP (False Positive) = 50: Incorrectly predicted default when there was no default
• FN (False Negative) = 50: Incorrectly predicted no default when there was default
• TP (True Positive) = 900: Correctly predicted default when there was default

To determine which value represents the correct predictions of defaults, we look at TP (True Positive) = 900, which corresponds to option D.

Key Metrics:

• Accuracy = (TP + TN) / Total = (900 + 100) / (100 + 50 + 50 + 900) = 1000 / 1100 ≈ 90.9%
• Precision = TP / (TP + FP) = 900 / (900 + 50) = 900 / 950 ≈ 94.7%
• Recall/Sensitivity = TP / (TP + FN) = 900 / (900 + 50) = 900 / 950 ≈ 94.7%
• Specificity = TN / (TN + FP) = 100 / (100 + 50) = 100 / 150 ≈ 66.7%

The model shows good performance in detecting defaults (high recall and precision) but has lower specificity in correctly identifying non-defaults.