Application of Particle Swarm Optimization Algorithm with Feed-Forward Neural Network for stock forecasting
The aim of this project is to compare ability of stock forecasting between simple Feed-Forward Neural Network and Feed-Forward Neural Network which using Particle Swarm Optimization to finding best value for Feed-Forward Neural Network weight.
- EMA 5 days
- EMA 10 days
- MACD
- RSI 14 days
- Feed-Forward Neural Network
- Particle Swarm Optimization
- Buy and Hold
- pyswarms Library (see detail here)
- create conda environment.
conda create -n myenv python=3.7
- activate your environment
conda activate myenv
- using pip to install require library
pip install -r requirements.txt
- running simulator (default simulate day is 30)
If you run the simulator in terminal you have to close the window that show the result graph for simulator next stock.
python run_simulator.py