Section 6.1. Data Mining Basic Concepts

In this section we explore the main concepts behind training and using a Machine Learning model to predict some kind of output; such as a behavior, a number, a class, a future time evolvement.

Machine Learning algorithms can cover many different types of applications, each requiring a specific type of model. For example, the most popular algorithms are supervised classification method, such as a decision tree or a logistic regression. An evolution of those involves the combination of many such models working in parallel. In case of unlabeled data sets, clustering algorithms might become useful to discover hidden similarities. For prediction over time, time series analysis techniques would be required.