Two broad families of algorithms will be discussed in this book:
In unsupervised learning, the algorithm will seek to find the structure that organizes unlabelled data. For instance, based on similarities or distances between observations, an unsupervised cluster analysis will determine groups and which observations fit best into each of the groups. An application of this is, for instance, document classification.
In supervised learning, we know the class or the level of some observations of a given target attribute. When performing a prediction, we use known relationships in labeled data (data for which we know what the class or level of the target attribute is) to predict the class or the level of the attribute in new cases (of which we do not know the value).
18.191.215.96