In this project we faced a classification problem on a very simple literature dataset, Monk, with the aim of becoming more familiar with libraries such as Keras, Tensorflow, PyTorch.
Then the real aim of the project was to find the best configuration of a neural network for the solution of a regression task with two outputs.
The project also saw the in-depth application of algorithms such as SVM and KNN, but focused in particular on neural networks, which were the main subject of the course.
Theory in this exam
Many hours were also dedicated to the good application of machine learning, to the theoretical study of the statistical theory behind the models.
Advanced topics such as reservoir computing, recurrent and recursive neural network were also seen, only at a theoretical level.
Download of the document
The complete report with all the details of the grid searches carried out, the validation techniques carried out, can be downloaded below.