Machine Learning · Time Series · Decision Support
Market Regime Detection & Backtesting Platform
Market conditions change how trading strategies behave. I built this platform to classify trend, range, and volatility regimes, then test strategy variants across them.
- 7
- strategy families
- 21
- strategy configurations
- Walk-forward
- out-of-sample validation
- What I built
- The Python workflow engineers time-series features, compares K-Means with an optional Gaussian HMM, and evaluates baseline, regime-aware, and liquidity-filtered strategies with next-bar execution, configurable costs, and Monte Carlo stress tests.
- Why it matters
- The dashboard keeps regime behavior, risk, and strategy comparisons in one place without presenting a single backtest as a final answer.
Pythonscikit-learnStreamlitPlotly
