michaeljf07/forex-predictor — reverse-engineered prompt

Build me a Python project that predicts next day forex price movement as an educational machine learning experiment, not a real trading bot. It should download recent market data from Yahoo Finance, including major currency pairs, stock indexes, commodities, bond yields, and VIX, then create lots of useful technical indicator features like moving averages, RSI, MACD, Bollinger Bands, volatility, momentum, and market correlations.

Train a few simple models, like Random Forest, Gradient Boosting, and Ridge Regression, compare how well they predict returns, and save the trained models and results. I want clear charts that show model performance, predictions versus actual prices, feature importance, residuals, and a simple strategy comparison against buy and hold.

Please make it easy to run from the command line with separate scripts for fetching data, building features and training models, and creating visualizations. Keep the code readable and beginner friendly, with comments where helpful, and include a clear disclaimer that this is only for learning and should not be used for real trading.