When and How to Apply Statistics, Machine Learning and Deep Learning Techniques

Machine Learning has become 'commodity' in engineering and experimental sciences, as calculus and statistics did before. After the hype produced during the 00's, machine learning (statistical learning, neural networks, etc.) has become a solid and reliable set of techniques available to the general researcher population to be included in their common procedures, far from the mysticism surrounding this field when only ML experts could solve modeling and prediction problems using such novel algorithms. But while knowledge on this field has settled among professionals, novice ML users still have trouble to decide when determined techniques could and should be applied to solve a given problem, sometimes ending with over-complicated solutions for simplistic problems, or complex problems partially solved by simplistic methods. This tutorial wants to introduce the most common techniques on statistical learning and neural networks, towards showing the proper techniques for each given scenario.