Deep learning for supervised classification

One of the most recent area in the Machine Learning research is Deep Learning. Deep Learning algorithms have been applied successfully to computer vision, automatic speech recognition, natural language processing, audio recognition and bioinformatics. The key idea of Deep Learning is to combine the best techniques from Machine Learning to build powerful general‑purpose learning algorithms. It is a mistake to identify Deep Neural Networks with Deep Learning Algorithms. Other approaches are possible, and in this paper we illustrate a generalization of Stacking which has very competitive performances. In particular, we show an application of this approach to a real classification problem, where a three-stages Stacking has proved to be very effective.