Bank Term Deposit Prediction
Machine learning solution for predicting bank term deposit subscriptions. Optimizes marketing campaigns through predictive analytics with comprehensive binary classification pipeline.
Business Impact
Problem Solved
Banks invest significant resources in marketing campaigns. This solution optimizes marketing spend by targeting high-probability prospects.
Value Delivered
Increases campaign conversion rates, improves customer experience, and maximizes return on investment for marketing activities.
Key Results
Comprehensive binary classification pipeline analyzing ~41,000 customers with multiple ML algorithms for actionable insights.
Dataset Information
Source
Bank Marketing Campaign Data
Size
~41,000 customers × 20+ features
Target
Term deposit subscription (yes/no)
Class Distribution
88% no subscription, 12% subscription
Feature Categories
Demographic
Age, job, marital status, education, default status
Financial
Housing loan, personal loan status
Campaign
Contact method, month, day, duration, contacts
Economic
Employment rate, CPI, confidence index
Model Comparison
Naive Bayes
Probabilistic- Fast training and prediction
- Good baseline performance
- Handles categorical features well
- Probabilistic outputs
Decision Tree (Gini)
Tree-based- Interpretable decision rules
- Handles mixed data types
- Provides feature importance
- No feature scaling required
Decision Tree (Entropy)
Tree-based- Alternative splitting criterion
- Comparable performance to Gini
- Different decision boundaries
- Information theory approach
Key Insights & Findings
Data Insights
Class Imbalance
~88% customers do not subscribe to term deposits
High Correlation
Strong correlation between economic indicators
Seasonal Patterns
March and December show higher subscription rates
Contact Method
Cellular contact generally outperforms telephone
Business Recommendations
Target Segments
Focus campaigns on students and retired individuals
Timing
Schedule major campaigns in March and December
Contact Method
Prioritize cellular over telephone contact
Previous Success
Heavily weight previous campaign success in targeting
Technical Architecture
Data Processing
Data loading, cleaning, comprehensive EDA with visualizations and data quality validation.
Feature Engineering
Creates meaningful features, handles multicollinearity, and implements categorical encoding.
Model Training
Implements and compares Naive Bayes and Decision Tree models with comprehensive evaluation.
Model Evaluation
Complete assessment with metrics, visualizations, confusion matrices, and ROC curves.
Model Persistence
Automated saving of trained models with timestamps and comprehensive reporting.