Support vector classifiers are one of the most popular linear classification techniques for binary classification. Different from some commonly seen model fitting criteria in statistics, such as the ordinary least squares criterion and the maximum likelihood method, its algorithm depends on an optimization problem under constraints, which is…

Over the past few years, academics have undertaken initiatives to bridge the gap between theory and practice in the ever-growing field of business analytics, including implementing real-life student projects in all shapes and forms. Every year since 2015, Manhattan College has invited student teams from across North America and elsewhere in the…

The methodology and mathematical treatment of several classic multivariate methods for the analysis of spectroscopic data is demonstrated in a straightforward way that can be used as a basis for teaching an undergraduate introductory course on chemometric analysis. The multivariate techniques of classical least-squares (CLS), principal component…