Short-Term SPY ETF Price Forecasting Model

Introduction

This project aims to predict the short-term high and low prices of the SPY ETF using various machine learning and statistical models. The primary objective is to aid retail traders, particularly day traders, in making informed investment decisions.

Data

The project utilizes historical price data and a suite of technical indicators as input for the models. Different Jupyter Notebooks come with their corresponding datasets.

Models

Several models were implemented and tested in this project:

• Baseline Model:

  • Average Model

• Advanced Models:

  • Holt
  • ARIMA
  • Linear Regression
  • Ridge Regression
  • Lasso Regression
  • Random Forest
  • Support Vector Regression (SVR)
  • Extreme Gradient Boosting (XGBoost)

• Advanced Time-Series Models:

  • Long Short-Term Memory (LSTM)
  • Hidden Markov Models (HMM)

Notebooks

• ML.ipynb: Use Advanced Models to predict High/Low for the next one hour for SPY.

• HMM.ipynb: Use HMM model to predict next one hour price for SPY.

• LSTM.ipynb: Train LSTM model to predict next one hour high price for SPY.

• Data.ipynb: Preproces and visualize some data.

• Data2.ipynb: Prepare train sets and test sets.

Authors

Team Algebros: Sailun Zhan, Xinwu Yang, Aolong Li, Amin Idelhaj, Zongze Liu

License

This project is licensed under the MIT License - see the LICENSE file for details.