Induction in Hierarchical Multi-label Domains

Induction of classifiers from sets of preclassified training examples is one of the most popular machine learning tasks. This book focuses on the techniques needed in the field of automated text categorization. Here, each document can be labeled with more than one class, sometimes with many classes. Moreover, the classes are hierarchically organized, the mutual relations being typically expressed in terms of a generalization tree. The new performance measure specifically designed for evaluating the classification performance of hierarchical classifiers is also presented. The new performance measure considers the underlying relationships among classses, and hence reflects the important properties of hierarchical classifiers accounting to the information inherent in the class hierarchy.

Autorentext

Sareewan Dendamrongvit received her Ph.D. degree in Electrical and Computer Engineering in 2011 from The University of Miami, Coral Gables, Florida, USA. Her main research interests include machine learning, data mining, and neural networks.