This book introduces the representative algorithms of machine learning in simple terms, reveals the statistical learning theory behind it; it is an introductory and advanced professional material for AI and machine learning. This book addresses the problem of supervised learning–one of the most common problems in machine learning–in three parts as introduction, advancement and deepening. All of the three parts involve basic introductory algorithms, core theories and mathematical optimization behind the theories. The introduction part uses the generalized linear model represented by logistic regression as the starting point and briefs on all the knowledge points involved in the book. In the advancement part, the core theories refer to empirical risk minimization, structural risk minimization, regularization and unified border theories. In the deepening part, the mathematical optimization includes the mathematical derivation of the maximum entropy principle, Lagrange duality and other theories, as well as the discussion of the mainstream optimization method for model solution.