A classifier operates on a vector of features, which are random variables, and outputs a decision as to which class the feature vector belongs. We consider binary classification, meaning that the decision is between two classes. Typically, a classifier is designed and its error estimated from sample data. Two basic questions arise. First, given a set of features, how does one design a classifier from the sample data that provides good classification over the general population? Second, how does one estimate the error of a designed classifier from the data? This book examines both classifier design and error estimation when one is given prior knowledge regarding the population. The first chapter provides basic background knowledge.
|