A quantitative finance research platform built from first principles to investigate option pricing, volatility forecasting, market regime detection, and decision-making under uncertainty.
This project explores how forecasting models, statistical learning techniques, and uncertainty quantification methods can be used to generate and evaluate volatility-based signals in financial markets.
The project combines:
- Option Pricing
- Volatility Forecasting
- Machine Learning
- Market Regime Detection
- Signal Generation
- Portfolio Construction
- Risk Evaluation
The broader goal is to study predictive modeling and decision-making under uncertainty, with applications in quantitative finance, operations research, and industrial engineering.
- Can option pricing models be implemented and validated from scratch?
- How accurately can volatility be estimated and forecasted?
- Do market volatility regimes naturally emerge from historical data?
- Does disagreement between forecast volatility and market expectations contain predictive information?
- Can volatility-based signals generalize across multiple assets?
-
Black-Scholes Pricing Model
-
European Call Pricing
-
European Put Pricing
-
Greeks
- Delta
- Gamma
- Vega
- Monte Carlo Simulation
- Finite Difference Validation
- Newton-Raphson Implied Volatility Solver
- Historical Volatility
- Volatility Smile Analysis
- IV vs HV Comparison
- GARCH(1,1) Forecasting
- Rule-Based Volatility Regimes
- K-Means Clustering
- Forecast Volatility vs Market Expectations
- Cross-Sectional Validation
- Confidence-Based Signal Filtering
- Equal Weight Portfolio
- Inverse Volatility Weighting
- Confidence-Based Position Sizing
| Metric | Value |
|---|---|
| Black-Scholes Price | 10.4506 |
| Monte Carlo Price | 10.4538 |
| Difference | 0.0033 |
| Metric | Value |
|---|---|
| Forecast Correlation | 0.9833 |
| Ticker | Accuracy |
|---|---|
| AAPL | 55.90% |
| MSFT | 52.81% |
| NVDA | 56.74% |
| AMZN | 53.09% |
| GOOG | 54.78% |
Average Accuracy: 54.66%
| Metric | Value |
|---|---|
| Accuracy | 56.18% |
| Precision | 56.11% |
| Recall | 56.74% |
| F1 Score | 56.42% |
| Threshold | Sharpe | Max Drawdown |
|---|---|---|
| 0% | -0.11 | -27.36% |
| 5% | 0.45 | -13.31% |
| 10% | 0.64 | -11.66% |
| Metric | Baseline | Optimized |
|---|---|---|
| Sharpe Ratio | -0.19 | -0.10 |
| Max Drawdown | -18.83% | -3.19% |
| Total Return | -4.47% | -0.44% |
Full report available:
research/report/Quant_Research_Report.pdf
- Python
- NumPy
- Pandas
- SciPy
- Scikit-Learn
- Matplotlib
- yFinance
- Historical Options IV Data
- Hidden Markov Models
- EGARCH and GJR-GARCH
- Volatility Surface Modeling
- Portfolio Optimization Under Uncertainty
- Stochastic Optimization
- Risk-Aware Decision Frameworks
Mehul Gupta
Master of Business Analytics, University of Illinois Chicago
Research Interests:
- Operations Research
- Quantitative Finance
- Machine Learning
- Optimization Under Uncertainty
- Decision Analytics