Description:
This case study explores using supervised decision tree models to predict personal loan purchases at AllLife Bank. It analyzes customer data to identify key factors influencing loan acceptance and offers insights for more effective marketing strategies.
Context:
AllLife Bank aims to convert depositors into loan borrowers to enhance its loan business and increase interest earnings. A past campaign showed a 9% conversion rate, prompting the need for more targeted marketing strategies.
Objective:
Predict whether a liability customer will purchase a personal loan.
Identify significant variables influencing loan acceptance.
Determine which customer segments to target.
Techniques Used:
Data Visualization: Seaborn and Matplotlib
Descriptive Analytics: Data summarization and interpretation
Machine Learning: Decision Tree Classifier, Logistic Regression
Model Tuning: GridSearchCV for hyperparameter optimization
Model Evaluation: Precision, Recall, F1-score, ROC-AUC
Data Dictionary:
Age: Customer’s age in years
Experience: Years of professional experience
Income: Annual income (in thousand dollars)
Family: Family size
CCAvg: Average monthly credit card spending (in thousand dollars)
Education: Education level (1: Undergrad; 2: Graduate; 3: Advanced/Professional)
Mortgage: Mortgage value (in thousand dollars)
Personal_Loan: Accepted the personal loan? (0: No, 1: Yes)
Securities_Account: Has a securities account? (0: No, 1: Yes)
CD_Account: Has a CD account? (0: No, 1: Yes)
Online: Uses internet banking? (0: No, 1: Yes)
CreditCard: Uses a credit card from another bank? (0: No, 1: Yes)
Domain:
Finance and Banking