Thanks to visit codestin.com
Credit goes to www.scribd.com

Scribd
Upload
444 views1 page

Task 2 Model Plan Example Answer

The document outlines a predictive model plan using logistic regression to estimate customer delinquency based on features like Credit Utilization and Missed Payments. It justifies the choice of logistic regression for its interpretability and efficiency in financial services, emphasizing the importance of model transparency. The evaluation strategy includes metrics such as accuracy and F1 score, with a focus on fairness and ethical considerations in model predictions.

Uploaded by

Vineeth Vivekanandan
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
444 views1 page

Task 2 Model Plan Example Answer

The document outlines a predictive model plan using logistic regression to estimate customer delinquency based on features like Credit Utilization and Missed Payments. It justifies the choice of logistic regression for its interpretability and efficiency in financial services, emphasizing the importance of model transparency. The evaluation strategy includes metrics such as accuracy and F1 score, with a focus on fairness and ethical considerations in model predictions.

Uploaded by

Vineeth Vivekanandan
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 1

Predictive Model Plan – Example Answer

1. Model Logic (Generated with GenAI)


Using ChatGPT, I generated a predictive model using logistic regression to estimate the
likelihood of a customer becoming delinquent. The model uses key features such as
Credit_Utilization, Missed_Payments, Income, Debt_to_Income_Ratio, and Account_Tenure to
predict a binary outcome: 1 if the customer is likely to become delinquent, and 0 otherwise.

Pseudo-code:
1. Load dataset
2. Select features: ['Credit_Utilization', 'Missed_Payments', 'Income',
'Debt_to_Income_Ratio', 'Account_Tenure']
3. Define target variable: 'Delinquent_Account'
4. Split data into training and testing sets
5. Fit logistic regression model
6. Predict and evaluate using classification metrics

2. Justification for Model Choice


I chose logistic regression because it is widely used for binary classification problems and is
highly interpretable. In financial services, model transparency is crucial, and logistic regression
offers clear coefficient outputs to explain each predictor’s influence.

The model is simple to implement, does not require large computational resources, and provides
strong baseline performance for credit risk analysis. It allows for quick iteration and stakeholder
communication, making it an ideal fit for Geldium’s goal of responsibly identifying at-risk
customers.

3. Evaluation Strategy
To evaluate the model, I would use accuracy, precision, recall, F1 score, and AUC. Precision and
recall are particularly important: precision ensures we avoid unnecessary interventions for low-
risk customers, while recall helps identify most high-risk customers.

The F1 score balances both, and AUC evaluates the model’s ability to distinguish between
classes across thresholds.

To check for bias, I would examine prediction patterns across demographic segments (e.g.,
Employment_Status or Location) to ensure fairness.

Any strong disparities would prompt model reassessment or rebalancing. Ethical considerations
include avoiding proxy bias, maintaining transparency, and clearly communicating how model
outputs influence decisions.

You might also like