- 5.8k Downloads
The goal of classification, also known as supervised pattern recognition, is to provide a model that yields the optimal discrimination between several classes in terms of predictive performance. It is closely related to clustering. The difierence is that in classification it is clear what to look for: the number of classes is known, and the classes themselves are well-defined, usually by means of a set of examples, the training set. Labels of objects in the training set are generally taken to be error-free, and are typically obtained from information other than the data we are going to use in the model. For instance, one may have data – say, concentration levels of several hundreds of proteins in blood – from two groups of people, healthy, and not-so-healthy, and the aim is to obtain a classification model that distinguishes between the two states on the basis of the protein levels.
KeywordsHide Layer Discriminant Analysis Linear Discriminant Analysis Gini Index Hide Unit
Unable to display preview. Download preview PDF.