Machine Learning with SVM and Other Kernal Methods

Support vector machines (SVMs) represent a breakthrough in the theory of learning systems. It is a new generation of learning algorithms based on recent advances in statistical learning theory. Designed for the undergraduate students of computer science and engineering, this book provides a comprehensive introduction to the state-of-the-art algorithm and techniques in this field. It covers most of the well known algorithms supplemented with code and data. One Class, Multiclass and hierarchical SVMs are included which will help the students to solve any pattern classification problems with ease and that too in Excel. This title includes extensive coverage of Lagrangian duality and iterative methods for optimization; separate chapters on kernel based spectral clustering, text mining, and other applications in computational linguistics and speech processing; a chapter on latest sequential minimization algorithms and its modifications to do online learning; step-by-step method of solving the SVM based classification problem in Excel; and, kernel versions of PCA, CCA and ICA. The CD accompanying this book includes animations on solving SVM training problem in Microsoft EXCEL and by using SVMLight software. In addition, Matlab codes are given for all the formulations of SVM along with the data sets mentioned in the exercise section of each chapter.

Table of Contents:
Preface; 1. Machine Learning with Support Vector Machines; 2. Supervised Automatic Learning: A Probabilistic Framework; 3. Essential Mathematical Background; 4. Kernel Methods and the Evolution of SVM; 5. Support Vector Regression; 6. Simple Variants of SVM: Mangasarian's Approaches; 7. Sequential Minimisation Algorithms; 8. One Class SVM; 9. Multi-class and Hierarchical Support Vector Machines; 10. String Kernels. 11. Kernel-based Methods for Clustering Data; 12. Data Sets; 13. Other Kernel Methods K-PCA, K-CCA, K-PLS, K-ICA; 14. Kernel Methods for Text Categorisation; 15. Kernel Methods for Speech Recognition; 16. Kernel Methods in Natural Language Processing: An Introduction; Appendix; Index.

About the Author :
K.P. SOMAN (M.Tech and Ph.D., IIT Kharagpur) is Head, Centre for Excellence in Computational Engineering and Networking, Amrita Vishwa Vidyapeetham, Coimbatore. He has published/presented over 80 papers in international journals and conferences. He has executed several funded projects for organizations/institutions like Indian Space Research Organization (ISRO), Naval Physical and Oceanographic Laboratory (NPOL), Cochin, Aeronautical Development Establishment (ADE) Bangalore and Ministry of Communication and Information Technology. His areas of interest include high performance computing, Machine learning, wavelet transform, fractal analysis and cryptography. Dr. Soman has authored two books-Insight into Wavelets: From Theory to Practice, and Insight into Data Mining: Theory and Practice-both published by PHI Learning, New Delhi. R. LOGANATHAN [B.E., M.S. (by Research)] is Research Associate, Centre for Excellence in Computational Engineering and Networking, Amrita Vishwa Vidyapeetham, Coimbatore. His areas of interest are machine learning applications to natural language processing, computational linguistics and machine translation. V. AJAY (B.E., M.Tech) is Research Scholar, Department of Mechanical Engineering, Purdue University, USA. He has co-authored the book Insight into Data Mining: Theory and Practice-published by PHI Learning. His areas of interest include nano-material simulation, computational linear algebra and optimization.