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 $m_1$ and $m_2$ based on their cartesian coordinates, using the initial conditions $z_0 = [π/4,0,π/4,0] $

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, $z_0= [π/2,0,π/2,0]$, leading to more intricate pendulum behaviors.

Selective Training: Explored predictions when the RNN is trained solely using the coordinates of mass $m_2$, without insights from mass $m_1$.

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.