unizg Search

Machine Learning

Teachers in charge

Prof. Tomislav Rolich, Ph.D.

Course code

230132

Study Module

  • TTI - graduate track - Engineering Design and Management of Textiles
  • TTI - graduate track - Textile Chemistry, Materials, and Ecology
  • TTI - graduate track - Clothing Engineering
  • TTI - graduate track - Textile Design for Industry
  • TTI - graduate track - Clothing Design for Industry

Schedule ∑ (P+V+S)

4 (2+2+0)

ECTS

4

Knowledge verification

Colloquium, written exam, oral exam

Precondition for testing

 

Lecture type

Lectures, exercises, seminars

Exercise type

Auditory, laboratory

Learning outcomes 
the students will be able to:

  • Explain the basic theoretical concepts of machine learning, the working principles of selected machine learning algorithms, and choose an appropriate algorithm to solve a given problem.
  • Apply machine learning algorithms to practical examples from the textile and clothing industry and other engineering fields using the Weka software system.
  • Analyze data sets and choose appropriate options in the Weka software program to solve a given problem.
  • Model their own algorithms suitable for solving a given problem on the provided data set using the Weka software system.
  • Interpret the results of the algorithms (efficiency and other parameters) for a specific data set in the Weka software system.

Subject content

Introduction to the basic concepts of machine learning and a brief overview of the most popular machine learning algorithms. Introduction to computer applications for solving machine learning problems. Introduction to the concept of training and testing sets, the importance of repeated experiments, and the method for determining classifier efficiency. Introduction to the concept of cross-validation and interpretation of the results obtained in this way. Introduction to the OneR algorithm, the concept of overfitting, using probability in the construction of classifiers (Naive Bayes algorithm). Creation of decision trees, use of the linear regression algorithm. Introduction to data mining. Introduction to the Experimenter interface in the Weka program. Discretization of numerical attributes. Introduction to simple neural networks, the multilayer perceptron, and learning curves.

Aim of course

To train students to apply machine learning algorithms on large data sets using the Weka software system. To train students to independently model machine learning algorithms for solving a given problem. To train students to independently draw conclusions based on the analysis of algorithm results and to propose the implementation of these conclusions in the decision-making process.

Literature necessary for the course

Supplementary literature

Machine Learning