Aine eesmärk on anda ülevaade masinõppest, seal kasutatavatest meetoditest ja printsiipidest, tulemuste hindamisest ja rakendatavusest, samuti arendada oskusi masinõppe mudelite koostamisel.

The goal is to give an overview of main principles and methods in machine learning. The applicability and assessment of machine learning models are handled as well. The skills of developing of machine learning models is an important goal as well.

Õppeaine läbinud üliõpilane:
- nimetab masinõppe põhimõisteid, -printsiipe, -meetodeid ja algoritme;
- hindab masinõppe mudelite kasutamise täpsust ja otstarbekust konkreetsete probleemide ning ülesannete lahendamisel;
- kasutab masinõppe-alast tarkvara konkreetsete probleemide lahendamisel.

After completing this course, the student:
- formulates the key concepts, principles, algorithms and methods of machine learning;
- evaluates exactness and reasonability of applying of machine learning models in case of particular problems;
- utilizes machine learning oriented software for developing models.

Ülevaade masinõppe põhiprintsiipidest, eesmärkidest ja rakendusvaldkondadest. Käsitletakse juhendatud ja juhendamata õpet ning stiimulõpet. Meetoditest ja lähenemisviisidest vaadeldakse klasterdamist, mõõtmelisuse vähendamist, otsustuspuude põhiseid meetodeid, tehisnärvivõrke, Bayesi statistika põhiseid meetodeid, tugivektormasinaid, k lähima naabri põhist meetodit. Õpitakse ka mudeleid hindama. Praktiseeritakse ka eeltreenitud närvivõrkude adapteerimist konkreetsetele probleemidele.

Overview of main principles, goals and scope of application of machine learning. Supervised, unsupervised and reinforcement learning are handled. Main methods like clustering, reduction of dimensionality, decision trees, artificial neural networks, Bayesian methods, support vector machines, kNN-methods are considered. The assessment of models is handled. Adapting of pretrained neural networks in case of particular problems is considered as well.

-

-

Mudelite koostamine ja häälestamine. Koduse uurimistöö koostamine.

Developing and tuning of models. Composing the research work.

- Russell, S.J and Norvig, P. Artificial intelligence: a modern approach, third edition, Prentice Hall. 2010. ISBN-13: 978-0-13-207148-2
- Christopher Bischop. Pattern Recognition and Machine Learning. ISBN-13: 978-0387310732. Springer 2011.
- Zielesny A. From Curve Fitting to Machine Learning: An Illustrative Guide to Scientific Data Analysis and Computational Intelligence. Springer 2018.

Toomas Lepikult, dotsent (IC - IT kolledž)