Learning Objectives:

• Master classification fundamentals in financial contexts
• Understand decision tree algorithms and their applications
• Learn model evaluation metrics for classification
• Apply classification models to financial problems
• Interpret and validate classification results

Core Resources:

1. Theoretical Foundations

• Classification and Regression Trees - Breiman et al.
• Credit Scoring Using Machine Learning - Hand & Henley
• Bank Rating Prediction - Altman & Rijken
• Regulatory Pressure and Fire Sales in the Corporate Bond Market- Ellul & Jotikasthira & Lundblad
• Fire Sales and Impediments to Liquidity Provision in the Corporate Bond Market- Wang & Zhang & Zhang

2. Technical Implementation

• Scikit-learn Decision Trees
• Classification Metrics
• Model Validation

3. Financial Applications

• Data Sources:
  • S&P Credit Ratings
  • Bond Ratings
  • Ratings vs Rating History
  • Lending Club Dataset
  • FRED Economic Data
• Code Samples:
  • Lending Club Code
• Applications:
  • Credit Risk Assessment
  • Rating Changes Prediction
  • Fraud Detection
  • Trading Signal Generation
  • Default Prediction

Weekly Assignment

Option 1: Credit Risk Classification

1. Data Preparation
   • Use Lending Club or WRDS data
   • Prepare features: credit scores, debt ratios, employment history
   • Handle class imbalance
   • Split data respecting time order
2. Model Implementation
   • Implement decision tree classifier
   • Use cross-validation for parameter tuning
   • Consider class weights or sampling techniques
   • Compare with logistic regression
3. Analysis
   • Evaluate using precision, recall, F1-score
   • Analyze feature importance
   • Visualize decision tree structure
   • Consider economic implications of misclassification

Option 2: Research-Based Application

Design your own classification application in finance. Some suggestions:

• Rating Change Prediction
  • Predict credit rating upgrades/downgrades
  • Use financial ratios and market indicators
  • Compare with rating agency decisions
• Trading Signal Generation
  • Classify market conditions
  • Generate buy/sell signals
  • Evaluate trading performance

Implementation Tips: