Predicting Monthly Excess Returns of CRSP Market Index

📌Overview

This project applies machine learning techniques( random forest & ridge ) to forecast monthly excess returns of the CRSP value-weighted market index using the Welch and Goyal financial indicators dataset. Covering data from 1927 to 2022, the analysis involves processing the dataset, calculating various financial indicators, and developing predictive models.

📚Key Libraries:

• pandas
• numpy
• scikit-learn
• seaborn
• matplotlib

🪜Key Steps:

📎Data Preparation:

Dataset Preview - Summary Statistics - Missing Data - Correlation Analysis - Visualization - Feature Distribution - Feature Engnieering - Data Splitting

📎Modeling:

Model Selection ( Ridge - Random Forest) - Time-series Validation - Hyperparameter Tuning - Regularization - Residual Analysis - Economic Interpretation

📎Evaluation:

Performance Metrics - Out-of-sample Performance - Visualization of Results - Comparative Analysis

📎Financial Analysis:

Key Predictor Analysis - Predictor Stability - Investment Strategies - Visualization of Insights