This repository contains the source code used for my undergraduate thesis.
Ensure you have Python installed on your system. TensorFlow 2 supports:
- Python: 3.8–3.11
- Operating Systems:
- Ubuntu 16.04 or later
- Windows 7 or later (with C++ redistributable)
- macOS 10.12.6 (Sierra) or later (no GPU support)
- WSL2 via Windows 10 19044+ (Experimental GPU support)
Verify Python installation:
python3 --version # python --version for WindowsIf Python is not installed, download it from the official website.
With the virtual environment activated, install the required packages:
pip install --upgrade pip
pip install -r requirements.txtMake sure to download the GPU drivers if not running on the cloud here
To scrape PSE stock data, run:
python scrape_data.pyThis script utilizes tickers.txt to determine which stocks to fetch and stores the data in the data/ directory.
The provided notebook runs the methodology outlined in the paper. It skips the data collection part since it is already collected via scrape_data.py, and ends with the backtesting simulations.
Run it via:
jupyter notebook notebook.ipynbThe repository organizes data as follows:
/pse-stocks
├── data/
└── predictions/
└── results/
data/: Contains scraped stock data from MarketWatch.predictions/: Contains actual vs predicted price plots for the final model.results/: Contains backtesting results for each strategies.
The project saves the final trained model to a file called model.keras, which can be loaded for other uses.