Machine Learning and Data Analysis

Description: The discipline is devoted to the study of algorithms, models and methods of machine learning and how to use them to solve practical problems; processing and analysis of big data; fundamentals of building models and algorithms for solving and evaluating the accuracy of solving problems using machine learning methods. Software tools, modern programming languages and libraries for analyzing real data sets are considered.

Amount of credits: 6

Пререквизиты:

• Linear Systems of Automatic Control

Course Workload:

Component: University component

Cycle: Profiling disciplines

Goal

• studying modern methods and tools of machine learning and data analysis to solve specific problems in the field of building automated and intelligent control systems for technical objects.

Objective

• formation of a holistic view of machine learning methods for processing and analyzing big data;
• mastering the skills of developing programs in Python and MatLab languages that implement machine learning algorithms;
• mastery of models and methods of data mining and machine learning in tasks of information retrieval, data processing and analysis;
• acquiring the skills of a data scientist and developer of mathematical models, methods and algorithms for data analysis.

Learning outcome: knowledge and understanding

• demonstrate knowledge of the mathematical foundations of machine learning theory, the main classes of machine learning algorithms.

Learning outcome: applying knowledge and understanding

• select the most appropriate algorithms for solving machine learning problems and evaluate the quality of the constructed models
• apply methods of data mining and machine learning in problems of information retrieval, processing and analysis of data from automation systems of technical objects.

Learning outcome: formation of judgments

• analyze, highlight features and combine machine learning methods;

Learning outcome: communicative abilities

• report trends in the field of automation of production systems, promising directions and opportunities for practical application, both to specialists and non-specialists;

Learning outcome: learning skills or learning abilities

• demonstrate the skills of a data scientist and developer of mathematical models, methods and algorithms for data analysis.

Teaching methods

modular learning technology

information and communication technologies

technologies of educational and research activities

Assessment of the student's knowledge

Teacher oversees various tasks related to ongoing assessment and determines students' current performance twice during each academic period. Ratings 1 and 2 are formulated based on the outcomes of this ongoing assessment. The student's learning achievements are assessed using a 100-point scale, and the final grades P1 and P2 are calculated as the average of their ongoing performance evaluations. The teacher evaluates the student's work throughout the academic period in alignment with the assignment submission schedule for the discipline. The assessment system may incorporate a mix of written and oral, group and individual formats.

The evaluating policy of learning outcomes by work type

Evaluation form

The student's final grade in the course is calculated on a 100 point grading scale, it includes:

• 40% of the examination result;
• 60% of current control result.

The final grade is calculated by the formula:

 

Where Midterm 1, Midterm 2are digital equivalents of the grades of Midterm 1 and 2;

E is a digital equivalent of the exam grade.

Final alphabetical grade and its equivalent in points:

The letter grading system for students' academic achievements, corresponding to the numerical equivalent on a four-point scale:

Topics of lectures

• Introduction to the theory and practice of machine learning
• Machine learning tasks
• Methods of teaching and assessing its quality
• Typical tasks when preparing data and training models
• Mathematical support for the theory of machine learning
• Machine learning models and methods
• Fundamental algorithms
• Anatomy of learning algorithms
• Feature design, retraining and hyperparameter tuning
• Models and methods of fuzzy logic
• Neural networks and deep learning
• Problems and solutions for training models
• Advanced techniques
• Learning without a teacher
• Development of applications in the field of machine learning

Key reading

• Burkov A. Machine learning without further ado. - St. Petersburg: Peter, 2020. - 192 p.
• Fundamentals of machine learning: textbook / O.V. Limanovskaya, T.I. Alferyeva; Ministry of Science and Higher Education Education of the Russian Federation. - Ekaterinburg: Ural Publishing House. University, 2020. - 88 p.
• Gladilin P.E., Bochenina K.O., Machine learning technologies – St. Petersburg: ITMO University, 2020. – 75 p.

Further reading

• Theory and practice of machine learning: textbook / V.V. Voronina, A.V. Mikheev, N.G. Yarushkina, K.V. Svyatov. – Ulyanovsk: Ulyanovsk State Technical University, 2017. – 290 p.
• Mehryar Mohri, Afshin Rostamizadeh, Ameet Talwalkar. Foundations of Machine Learning. - MIT Press, 2018. – 504с.