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 ordinary least squares criterion and maximum likelihood method, its algorithm depends on an optimization problem under constraints, which is unconventional to many students a second or third course stat...