Hybrid Intrusion Detection

Newer intrusions are coming out every day with the all-way growth of the Internet. In this context, this book proposes a hybrid approach of intrusion detection along with architecture. The proposed architecture is flexible enough to carry intrusion detection tasks either by using a single module or by using multiple modules. Two modules - (1) Clustering-Outlier detection followed by SVM classification and (2) Incremental SVM with Half-partition method, are proposed in the book. Firstly, this work develops the "Clustering-Outlier Detection" algorithm that combines k-Medoids clustering and Outlier analysis. Secondly, this book introduces the Half-partition strategy and also designs "Candidate Support Vector Selection" algorithm for incremental SVM. This book is intended for the people who are working in the field of Intrusion Detection and Data Mining. Researchers and Scholars who are interested in k-Means and k-Medoids clustering and SVM classification in particular, will find this book useful. Students who want to pursue their research work in the fields of Information Security and Data Mining may also consider this as a good reference.