1st Edition

Support Vector Machines and Their Application in Chemistry and Biotechnology

    211 Pages 70 B/W Illustrations
    by CRC Press

    212 Pages 70 B/W Illustrations
    by CRC Press

    Support vector machines (SVMs) are used in a range of applications, including drug design, food quality control, metabolic fingerprint analysis, and microarray data-based cancer classification. While most mathematicians are well-versed in the distinctive features and empirical performance of SVMs, many chemists and biologists are not as familiar with what they are and how they work. Presenting a clear bridge between theory and application, Support Vector Machines and Their Application in Chemistry and Biotechnology provides a thorough description of the mechanism of SVMs from the point of view of chemists and biologists, enabling them to solve difficult problems with the help of these powerful tools.

    Topics discussed include:

    • Background and key elements of support vector machines and applications in chemistry and biotechnology
    • Elements and algorithms of support vector classification (SVC) and support vector regression (SVR) machines, along with discussion of simulated datasets
    • The kernel function for solving nonlinear problems by using a simple linear transformation method
    • Ensemble learning of support vector machines
    • Applications of support vector machines to near-infrared data
    • Support vector machines and quantitative structure-activity/property relationship (QSAR/QSPR)
    • Quality control of traditional Chinese medicine by means of the chromatography fingerprint technique
    • The use of support vector machines in exploring the biological data produced in OMICS study

    Beneficial for chemical data analysis and the modeling of complex physic-chemical and biological systems, support vector machines show promise in a myriad of areas. This book enables non-mathematicians to understand the potential of SVMs and utilize them in a host of applications.

    Overview of support vector machines
    Maximal Interval Linear Classifier
    Kernel Functions and Kernel Matrix
    Optimization Theory
    Elements of Support Vector Machines
    Applications of Support Vector Machines

    Support vector machines for classification and regression
    Kernel Functions and Dimension Superiority
    Notion of Kernel Functions
    Kernel Matrix
    Support Vector Machines for Classification
    Computing SVMs for Linearly Separable Case
    Computing SVMs for Linearly Inseparable Case
    Application of SVC to Simulated Data
    Support Vector Machines for Regression
    ε-Band and ε-Insensitive Loss Function
    Linear ε-SVR
    Kernel-Based ε-SVR
    Application of SVR to Simulated Data
    Parametric Optimization for Support Vector Machines
    Variable Selection for Support Vector Machines
    Related Materials and Comments
    VC Dimension
    Kernel Functions and Quadratic Programming
    Dimension Increasing versus Dimension Reducing
    Appendix A: Computation of Slack Variable-Based SVMs
    Appendix B: Computation of Linear ε-SVR

    Kernel methods
    Kernel Methods: Three Key Ingredients
    Primal and Dual Forms
    Nonlinear Mapping
    Kernel Function and Kernel Matrix
    Modularity of Kernel Methods
    Kernel Principal Component Analysis
    Kernel Partial Least Squares
    Kernel Fisher Discriminant Analysis
    Relationship between Kernel Function and SVMs
    Kernel Matrix Pretreatment
    Internet Resources

    Ensemble learning of support vector machines
    Ensemble Learning
    Idea of Ensemble Learning
    Diversity of Ensemble Learning
    Bagging Support Vector Machines
    Boosting Support Vector Machines
    Boosting: A Simple Example
    Boosting SVMs for Classification
    Boosting SVMs for Regression
    Further Consideration

    Support vector machines applied to near-infrared spectroscopy
    Near-Infrared Spectroscopy
    Support Vector Machines for Classification of
    Near-Infrared Data
    Recognition of Blended Vinegar Based on
    Near-Infrared Spectroscopy
    Related Work on Support Vector Classification on NIR
    Support Vector Machines for Quantitative Analysis of
    Near-Infrared Data
    Correlating Diesel Boiling Points with NIR Spectra
    Using SVR
    Related Work on Support Vector Regression on NIR
    Some Comments

    Support vector machines and QSAR/QSPR
    Quantitative Structure-Activity/Property Relationship
    History of QSAR/QSPR and Molecular Descriptors
    Principles for QSAR Modeling
    Related QSAR/QSPR Studies Using SVMs
    Support Vector Machines for Regression
    Dataset Description
    Molecular Modeling and Descriptor Calculation
    Feature Selection Using a Generalized
    Cross-Validation Program
    Model Internal Validation
    PLS Regression Model
    BPN Regression Model
    SVR Model
    Applicability Domain and External Validation
    Model Interpretation
    Support Vector Machines for Classification
    Two-Step Algorithm: KPCA Plus LSVM
    Dataset Description
    Performance Evaluation
    Effects of Model Parameters
    Prediction Results for Three SAR Datasets

    Support vector machines applied to traditional Chinese medicine
    Traditional Chinese Medicines and Their Quality Control
    Recognition of Authentic PCR and PCRV Using SVM
    Data Description
    Recognition of Authentic PCR and PCRV Using
    Whole Chromatography
    Variable Selection Improves Performance of SVM
    Some Remarks

    Support vector machines applied to OMICS study
    A Brief Description of OMICS Study
    Support Vector Machines in Genomics
    Support Vector Machines for Identifying Proteotypic
    Peptides in Proteomics
    Biomarker Discovery in Metabolomics Using Support
    Vector Machines
    Some Remarks



    Yizeng Liang and Qing-Song Xu are with Central South University in Changsha, China.