Diederich J. (Ed.) Rule Extraction from Support Vector Machines

Berlin: Springer, 2008. - 264 p.
A significant barrier to the widespread application of support vector machines is the absence of a capability to explain, in a human comprehensible form, either the process by which an SVM arrives at a specific decision/result, or more general, the totality of knowledge embedded in these systems. This lack of a capacity to provide an explanation is an obstacle to a more general acceptance of back box machine learning systems. In safety-critical or medical applications, an explanation capability is an absolute requirement.
This book provides an introduction and overview of methods used for rule extraction from support vector machines. The first part offers an introduction to the topic as well as a summary of current research issues. The second chapter surveys the field of rule extraction from SVMs, reviews areas of current research and introduces an application in the financial field.