Classification can be performed using Artifical Neural Networks (ANNs), which categorize products into classes by detecting a number of separate features.
The goal is to load the data, divide it into training sets and use it to train a neural network. A feed-forward neural network with a back-propagation training algorithm will be used to recognize the classes of the unknown samples, with sigmoid and relu activation functions.