Author Description

This book deals with the task of prototype selection from the training set in the context of supervised learning. When the training set is large,use of the entire training set is often time-consuming and wasteful. In such cases, reducing the training set could be quite helpful. Prototype selection is also useful when the training set is too large for the problem at hand due to redundancies and unnecessary samples. In this case, the use of prototype selection can result in a higher classification accuracy. Prototype selection either means selecting a subset of the training set or selecting some prototypes which have been derived from the training set. The selection of prototypes entails answering questions, such as, how many prototypes to choose, whether to choose a subset of training patterns or not, which prototypes to choose, etc. These issues have been addressed in this book.