Machine learning
0800-UMASZ
1. Definitions of machine learning problem types.
2. Learning algorithm. What is it?
3. Types of learning. Learning from data.
4. Linear discrimination algorithms.
5. Regression algorithms.
6. Machine learning algorithms and Bayesian learning.
7. Parzen windows or NRBF network.
8. kNN method and some variants.
9. Decision trees algorithms.
10. SVM algorithm.
11. Ensembles
12. Some selection algorithms (feature, prototypes).
13. Clustering algorithms.
14. Methods of learning validation.
15. Meta-learning.
16. Analysis of data mining systems.
17. Usage of data mining systems.
Total student workload
30H - lecture
30H - laboratory/projects
90H - individual work
Learning outcomes - knowledge
Ability to differentiate between kinds of machine learning problems and data mining.
Student is able to prepare analysis of efficiency of learning phase.
Knowledge about linear and nonlinear discriminant analysis.
Knowledge about linear and non linear regression.
Knowledge about data mining systems.
Knowledge about clustering methods, prototype selection and feature selection.
W 2, 6
Knowledge about advanced construction and analysis of algorithms, optimization methods. Knowledge about numerical methods.
Knowledge about specialization of study truck.
Learning outcomes - skills
Ability to understand a task as the machine learning problem.
Is able to recognize between different kinds of learning problems.
Student is able to use different algorithms of discrimination and regression.
Understanding of fault usage of machine learning methods.
Student is able to use data mining systems.
U 1, 3, 6, 7, 8, 9, 10,11
Student efficiently search information about solving computer science problems. Student search and use knowledge about computer science methods (and some familiar methods as well).
Student has ability to individual work, is able to define necessary knowledge scope to realize some tasks. Student is able to self learning, to individual search of appropriate methods to solve given problems (using different information sources).
Student is able to estimate usefulness of new methods around computer science, including diagnostic tools, and is able to choose best tools in given context.
Student is able to extend or rebuild computer science project, increase their efficiency using better/new algorithms.
Student is able to reproject solution by exchange of computer science infrastructure.
Student identifies source of problems. Student is able to specifications of solutions/projects (specifications of software, specifications of hardware,of computer networks etc.).
Student has ability of programing resources and data bases usage.
Student is able to collect appropriate computer science tools to solve problems.
Learning outcomes - social competencies
Student is able to use new knowledge in application to some social problems.
K 2, 5
Student is able to be creative in projecting of computer science applications.
Student can be successful in social or scientific project creation.
Teaching methods
Lecture: oral presentation of knowledge, discussion.
Exercises: discussions about tasks. Group or individual problem solving.
Case analysis.
Laboratory: Individual and group working + expert help. Discussion about given problems.
Observation/demonstration teaching methods
- simulation (simulation games)
- display
Expository teaching methods
- problem-based lecture
Exploratory teaching methods
- classic problem-solving
- brainstorming
- points system
- practical
- project work
- observation
- case study
- seminar
- laboratory
- experimental
- situational
- presentation of a paper
Online teaching methods
- methods referring to authentic or fictitious situations
- content-presentation-oriented methods
- methods developing reflexive thinking
- exchange and discussion methods
- cooperation-based methods
- integrative methods
- evaluative methods
Type of course
elective course
Prerequisites
Mathematical analysis, Algebra, Algorithms and data structures. Object oriented programming
Course coordinators
Bibliography
1. Duda, Hart, Stork, Pattern Classification, Wiley, 2001
2. D. Larose, Odkrywanie wiedzy z danych, PWN, 2006
3. T. Mitchell. Machine learning. McGraw Hill, 1997.
4. T. Hastie, R. Tibshirani, and J. Friedman. The Elements of Statistical Learning: Data Mining, Inference, and Prediction. Springer Series in Statistics. Springer, 2001.
5. V. Vapnik, The Nature of Statistical Learning Theory, Springer, 1995
6. B. Schoelkopf and A. Smola, Learning with Kernels, MIT, 2002
7. I. Guyon, S Gun, M. Nikravesh, L Zadeh ed., Feature extraction, foundations and Applications, Springer, 2006
8. N. Jankowski, Ontogeniczne sieci neuronowe. O sieciach zmieniających swoją strukturę, Exit, 2003
9. V. Cherkassky and F. Mulier, Learning from data, Wiley, 1998
Additional information
Additional information (registration calendar, class conductors,
localization and schedules of classes), might be available in the USOSweb system: