Classifiers are 'black boxes' that examine an input and place the input into a category. Examples include medical diagnostics (presence/absence of a disease), quality control (is a part defective?), search engine results (does a page match the search criteria?), language identification (what is the language of a document?), and data quality metrics (does this data set have sufficient quality metrics?). Increasingly, software tools are developed that implement or automate the classification process. Competing classification algorithms are compared to determine which is better suited for a particular application. This book develops the tools needed to measure classifier performance, compare classifiers, and rank order the results. This text examines the properties of binary and multiclass classifiers from a general perspective. Topics include classifier performance metrics, error analysis, comparison of classifier performance, metric distributions, and receiver operator characteristics (ROC).