Statistical classification analysis
Course level: Postgraduate level
Number of credits: 8 hp
Course literature: "Discriminant Analysis and Statistical Pattern Recognition" by G.J. McLachlan (2004) and "Statistical Regression and Classification - From Linear Models to Machine Learning" by N. Matloff (2017), as well as articles if needed.
Course content and learning objectives: Likelihood-Based Approaches to classification, Classification via Normal models, Linear and quadratic classifiers, Classification using logistic models, Non-parametric classification, Misclassification error.
After completing the course, the student should be able to:
- explain and formulate the theoretical concepts important for linear and quadratic classification, as well as logistic regression;
- understand and use non-parametric classification methods;
- understand the limitations of the different classification methods;
- calculate, interpret and evaluate probabilities of misclassification;
- identify the strengths and weaknesses of different statistical classifiers and use them in practice;
- implement statistical classifiers using statistical software and draw adequate conclusions.
Organisation: Lectures, projects with presentations, and home assignments.
Examination: Home assignments and projects with presentations.
Prerequisites: Elementary multivariate normal distribution theory, statistical regression analysis.
Grading scale: Pass/Failed
Language of instruction: English
Course web page (in Lisam)
Last updated: 2022-01-19