Question 4
Milo’s Evaluation (ME) is a loan default prediction company that uses manual workforce to suggest possible defaulters. Dr. Milo, the Chief Data Officer of ME, has decided to replace the manual workforce with a cutting-edge Logistic Regression Model that takes “Age” and “Income” as parameters to predict defaulters. The model is configured such that a “Defaulter” is categorised as the positive class. The final model had the parameters as specified in Table-1 below. To test the model, the data in Table-2 is used as test data. Given this information, answer the given subquestions.
| Parameter (Independent variable) | Corresponding coefficient value (Beta value) |
|---|---|
| Intercept | +3.2 |
| Age (in tens of years) | -1.7 |
| Income (in Lakhs of INR per month) | +1.4 |
Table-1
| Customer ID | Age (in tens of years) | Income (in Lakhs INR per month) | Past Loan Repayment Status |
|---|---|---|---|
| CID01 | 4.0 | 0.8 | DEFAULTER |
| CID02 | 5.3 | 3.0 | DEFAULTER |
| CID03 | 2.7 | 2.5 | NOT DEFAULTER |
| CID04 | 3.5 | 2.8 | NOT DEFAULTER |
| CID05 | 4.7 | 2.0 | NOT DEFAULTER |
Table-2
At a threshold of 0.8, how many “False Negatives” is the model predicting?