A Neural Network working with a logistic regression algorithm and gradient descent, used to classify images from Molecular Dynamics and differentiate between solid and gas+liquid states.

The simulations were obtained with a C code inside the repository using the Andersen thermostat algorithm.

The images were obtained using VMD.

Part of the code used in the Neural Network comes from a Coursera graded assignment on logistic regression inside the Neural Network and Deep Learning specialization, teached by professor Andrew Ng.

The knowledge needed to perform the molecular simulations and much more come from my course of Laboratory of Computational Physics from Università degli studi di Palermo, my teachers were prof. Grazia Cottone and prof. Salvatore Miccichè. Both of them have my gratitude.