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

Scribd
Upload
7 views2 pages

Document 9

The document outlines a predictive model plan for assessing customer delinquency based on financial data and behavior, utilizing a Logistic Regression model. It details the step-by-step process for data preparation, model training, and evaluation metrics such as accuracy and recall. Additionally, it emphasizes the importance of ethical considerations and bias checks in the model's implementation.

Uploaded by

kk595139
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
7 views2 pages

Document 9

The document outlines a predictive model plan for assessing customer delinquency based on financial data and behavior, utilizing a Logistic Regression model. It details the step-by-step process for data preparation, model training, and evaluation metrics such as accuracy and recall. Additionally, it emphasizes the importance of ethical considerations and bias checks in the model's implementation.

Uploaded by

kk595139
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

Predictive Model Plan – Student Template

1. Model Logic (Generated with GenAI)


What the model does:
This model predicts whether a customer is likely to become delinquent (i.e., miss
future payments or already has a delinquent account) based on their financial data,
credit history, and account behavior.

Step-by-Step Process (in simple words):

1. Get the data ready:


a. Fill in any missing values (like missing income).
b. Change words (like job type or city) into numbers so the computer can
use them.
c. Combine 6 months of payment history into one simple feature (like total
missed payments).
2. Split the data:
a. Use 80% of the data to train the model.
b. Use 20% to test how well it works.
3. Build the model:
a. Use a Logistic Regression model because it's good for yes/no
predictions (delinquent or not).
4. Check how good the model is:
a. Use things like accuracy and recall to see if it predicts well.
b. Try to fix it if it gives too many wrong results.

✅ 2. Justification for Model Choice


Chosen Model: Logistic Regression

Why this model?

• It's easy to understand — shows which things (like income or missed


payments) affect the result the most.
• It's fast and works well with this type of data.
• It's transparent — the company (like Geldium) can explain the results to
managers or customers.
• It works well for yes/no problems — like predicting if a person will be delinquent
(1) or not (0).
• It's commonly used in finance for risk and fraud predictions.

✅ 3. Evaluation Strategy

How to check if the model is good:

• Accuracy: Did it predict right overall?


• Precision: When it said “delinquent,” was it usually correct?
• Recall: Did it catch most of the real delinquents?
• F1 Score: A mix of precision and recall.
• AUC Score: How well it separates the two groups (delinquent vs. not).

Bias Checks & Ethics:

• If too many customers are "not delinquent", we will balance the data (so both
classes are treated fairly).
• Make sure the model doesn’t unfairly treat people based on job, location, or
income.
• Avoid using sensitive or private data.
• Always review high-risk predictions manually before taking action.
• Update the model often, so it stays fair and accurate.

You might also like