Acceptance rates
Setting an acceptance rate and calculating the threshold for that rate can be used to set the percentage of new loans you want to accept. For this exercise, assume the test data is a fresh batch of new loans. You will need to use the quantile()
function from numpy
to calculate the threshold.
The threshold should be used to assign new loan_status
values. Does the number of defaults and non-defaults in the data change?
The trained model clf_gbt
and the data frame of it's predictions, test_pred_df
, are available.
This exercise is part of the course
Credit Risk Modeling in Python
Exercise instructions
- Print the summary statistics of
prob_default
within the data frame of predictions using.describe()
. - Calculate the threshold for a
85%
acceptance rate usingquantile()
and store it asthreshold_85
. - Create a new column called
pred_loan_status
based onthreshold_85
. - Print the value counts of the new values in
pred_loan_status
.
Hands-on interactive exercise
Have a go at this exercise by completing this sample code.
# Check the statistics of the probabilities of default
print(____[____].describe())
# Calculate the threshold for a 85% acceptance rate
____ = np.____(____['prob_default'], ____)
# Apply acceptance rate threshold
____[____] = ____[____].apply(lambda x: 1 if x > ____ else 0)
# Print the counts of loan status after the threshold
print(____[____].____())