Skip to content

Implementation of a FeedForward Neural Network (FFNN) using Matlab. The network is trained using Gradient Descent in combination with the BackPropagation technique.

Notifications You must be signed in to change notification settings

anferico/ffneuralnet

Repository files navigation

FeedForward Neural Network

Implementation of a FeedForward Neural Network (FFNN) using Matlab. The network is trained using Gradient Descent combined with the BackPropagation technique, and its performances can be evaluated on two datasets, namely the Monk dataset (https://archive.ics.uci.edu/ml/datasets/MONK's+Problems) and a dataset of sensor readings.

The file neural_network.m carries out the whole training process (also plotting the learning curves and saving them to a file) given a set of hyperparameters. If you want to perform a grid search to do hyperparametrs tuning, run main.m instead.

The Datasets directory contains a couple of dataset that can be used to train the network and evaluate its performances.

About

Implementation of a FeedForward Neural Network (FFNN) using Matlab. The network is trained using Gradient Descent in combination with the BackPropagation technique.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages