Machine Learning Technologies

Major: System Analysis
Code of subject: 6.124.03.E.109
Credits: 6.00
Department: Information Systems and Networks
Lecturer: Ph.D., prof. Lytvyn Vasyl Volodymyrovych
Semester: 7 семестр
Mode of study: денна
Learning outcomes: • Principles of data preparation tasks for machine learning; • Models and methods of machine learning; • methods of quality assessment model.
Required prior and related subjects: • Mathematical Statistics • Theory of probability •Discrete Mathematic
Summary of the subject: Problem machine learning. Objects and features. Types of scales: binary, nominal, ordinal, quantitative. Methods of selecting attributes and methods of data preparation. Classes of problems: classification, regression, prediction, clustering. Concepts: model algorithms, teaching method, function loss and functional qualities, the principle of minimizing the empirical risk, sliding control. The problems of classification: Bayesian classification algorithms. Metric classification methods: the method of nearest neighbors and its generalizations, window of Parzen and potential function, selection standards and optimization metrics. Linear classification methods: logistic regression, support vector method, linear Perceptron. Tasks recovery regression, least squares method, linear and nonlinear regression, principal components method. Neural networks: a multi-layered structure of neural network, activation function, completeness dual-layer networks in space Boolean functions, the algorithm back-propagation errors, methods of optimizing the structure of the network. The task of clustering: types of cluster structures, graph clustering methods, hierarchical clustering, statistical clustering methods - EM-algorithm, k-means method. Logical classification methods, decision trees, algorithm ID3, weighing voting. Composition classifications: bustinh, behinh mixes algorithms. Search associative rules.
Assessment methods and criteria: • Current control (40%): written reports on laboratory work, essay, oral examination; • Final control (60% of exam): in written, verbally.
Recommended books: • Захарія Л.М. “Інформаційний пошук. Алгоритми класифікації текстових документів” методичні вказівки до дисципліни “Машинне навчання” Львів: Видавництво Національного університету “Львівська політехніка”, 2012. — 36 с. • Воронцов К.В.. Курс лекций Математические методы обучения по прецедентам, МФТИ, 2004—2008. Електронний ресурс. Режим доступу: www.ccas.ru/voron/teaching.html • Николенко С.И. Курс лекций по машинному обучению – слайды. Електронний ресурс. Режим доступу: http://logic.pdmi.ras.ru/~sergei/index.php?page=mlaptu09 • Дьяконов А.Г. Анализ данных, обучение по прецедентам, логические игры, системы WEKA, RapidMiner и MatLab. Учебное пособие. Електронний ресурс. Режим доступу: www.machinelearning.ru/wiki/images/7/7e/Dj2010up.pdf