HF_README_SVR.md

metadata

tags:
  - time-series
  - regression
  - svr
  - stock-prediction
  - technical-analysis
  - scikit-learn

SVR Model for AAPL Price Prediction (Technical Indicators)

This repository hosts a trained Support Vector Regression (SVR) model and its necessary preprocessing components (StandardScaler) for predicting the closing price of AAPL.

Model Details

• Algorithm: Support Vector Regression (SVR) with RBF/Linear Kernel (Tuned by Grid Search)
• Features: 37 features derived from technical analysis (SMA, Volatility, Returns) with lookbacks up to 252 days.
• Target: Next day's closing price.
• Training Period: 2023-01-01 to 2024-12-31

Inference

To use this model, you must correctly calculate and input all 37 technical features (including moving averages and volatility ratios) for the day prior to the prediction.

1. Load the svr_model.joblib and standard_scaler.joblib.
2. Calculate the 37 features for day $T$.
3. Scale the 37 features using the loaded StandardScaler.
4. Run the prediction.