Linear regression, loss functions in regression. Generalizing ability, gradient learning methods. Gradient methods, regularization. Linear classification, classification quality metrics, logistic regression. Polynomial regression. Approximation by polynomials of high degree. Multiclass classification. decision trees. Bagging, random forests and decomposing the error into bias and spread, etc.
Study of machine learning methods, learning technologies and building models for various applied tasks using Python libraries.
LO 1. Know the basics of machine learning; basics of building linear models of applied problems; Fundamentals of Python programming and its libraries.
LO 2. Be able to apply modern modeling techniques, regularization of overtrained models that can be used to include regularization parameters.
LO 3. Own training models for linear processes on polynomial features; software implementation of specific tasks of an applied nature.