Overview

The ongoing digital transformation is changing the way companies work and evolve. Improvements in the automatic acquisition of data from company operations have made automatic techniques for processing large volumes of information even more relevant. This micro-module aims to provide knowledge of the most common automatic learning techniques used by machines and best practices in designing solutions, measuring their performance, and understanding their potential and limitations.

Objectives

At the beginning of the micro-module, the basic concepts associated with Machine Learning will be addressed. The most common models/algorithms will then be explored, with the final part of the training focusing on the practical aspects of the correct implementation and performance measurement of the chosen approach.

Learning Outcomes

At the end of this module, the learner should be able to:

• Compare the current limitations of the most common machine learning techniques;
• Identify challenges where machine learning can help improve the performance of the solution;
• Integrate machine learning algorithms into problem-solving;
• Examine whether a given technique is producing good results;
• Choose and explain which techniques are appropriate for a given challenge.

Content

Module 1: Type of Learning

• Supervised
• Unsupervised
• Semi-supervised

Module 2: Learning Algorithms

• Linear Regression
• Naive Bayes
• Decision Trees
• kNN
• SVM
• Neural Networks

Module 3: Performance Measurement and Parameter Tuning

• Holdout, k-Fold and Leave-One-Out Methods
• Confusion Matrix
• Pearson Correlation