This project uses machine learning to guess the movement of a two-part pendulum system using limited data. The goal is to predict its unpredictable swings.
Design and Training of RNN: Utilized a Recurrent Neural Network to forecast the future positions of masses
Stability Analysis: Investigated how initial condition variations influence the RNN's performance.
Prediction Extent: Assessed the network's capability to predict long-term future positions.
Complex Path Prediction: Implemented the above steps but with a different set of initial conditions,
Selective Training: Explored predictions when the RNN is trained solely using the coordinates of mass
Ensure you have the necessary libraries and tools installed, including:
Jupyter Notebook or any Python environment.
Libraries: numpy, matplotlib, scipy, tensorflow (or your preferred ML library).
Go to a directory where you want to clone the repository and run:
git clone https://github.com/elilouise/Double-Pendulum-Prediction-RNN.git
Launch Jupyter Notebook. Open the main notebook and follow along with the sections and comments for a guided experience. Feel free to modify parameters and model architectures to improve predictions.