ABSTRACT
Classification is a fundamental activity of science (Barlow, 1991). The advancement of science depends on systems to label observations and experiences. This chapter focuses on a particular statistical tool that is available to evaluate classification rules and the prediction rules that may come from them. The most basic case of classification and prediction occurs when we have one piece of information on a set of observations to use to classify or predict, and that piece of information is binary (presence/absence, male/female, high/low, etc.). With one dichotomous piece of information in combination with the knowledge of which of two classes these observations belong, a 2×2 matrix (often called a confusion matrix) can be formed, in which the binary classification or prediction variable can be represented as the row of the matrix and the classes to which the observations are to be sorted are in the columns. An example of a 2×2 confusion matrix is presented in Table 3.1. Hypothetical example of a 2×2 confusion matrix. https://www.niso.org/standards/z39-96/ns/oasis-exchange/table">
Problem
Present
Absent
Screen
Present
a
b
c
d
Absent
