Informatics and Machine Learning

10Classification and Clustering

Numerous prior book, journal, and patent publications by the author [1–68] are drawn upon extensively throughout this chapter. Almost all of the journal publications are open access. These publications can typically be found online at either the author’s personal website (www.meta‐logos.com) or with one of the following online publishers: www.m‐hikari.com or bmcbioinformatics.biomedcentral.com.

A classifier is typically a simple rule whereby a class determination can be made, such as a decision boundary. Learning the decision rule, or a sufficiently good decision rule, especially if simple (and elegant), is the implementation aspect of a classifier, and can be difficult and time consuming. Even so, this is usually manageable because at least you have data to “learn from,” e.g. supervised learning, where you have instances and their classifications (or “labels”). Learning for classification can be done very effectively using Support Vector Machines (SVMs), as will be described in what follows. With clustering efforts, or unsupervised learning, on the other hand, we do not have the label information during training. In what follows SVMs will also be shown to be incredibly effective at clustering when used with metaheuristics to recover label information in a bootstrap learning process. Also shown will be implementation details for distributed SVM training, and other speedup optimizations, for practical deployment of the powerful SVM classification ...