| Course Unit Code | 155-1305/03 |
|---|
| Number of ECTS Credits Allocated | 6 ECTS credits |
|---|
| Type of Course Unit * | Compulsory |
|---|
| Level of Course Unit * | Second Cycle |
|---|
| Year of Study * | First Year |
|---|
| Semester when the Course Unit is delivered | Winter Semester |
|---|
| Mode of Delivery | Face-to-face |
|---|
| Language of Instruction | Czech |
|---|
| Prerequisites and Co-Requisites | There are no prerequisites or co-requisites for this course unit |
|---|
| Name of Lecturer(s) | Personal ID | Name |
|---|
| LAN177 | RNDr. Miroslav Langer, Ph.D. |
| HUD0118 | doc. Dr. Ing. Miroslav Hudec |
| Summary |
|---|
| The aim of the course is to understand and use stochastic and intelligent SC methods in economics for modeling and construction of flash predictions for economic and financial processes. These methods are based on supervised, unsupervised and hybrid learning from data, development of novel ANN architectures and design of novel systems for business applications. Students will be able to discuss and evaluate the performance of intelligent information processing in comparison with probabilistic computation. |
| Learning Outcomes of the Course Unit |
|---|
1. To gain a basic knowledge of SC information technologies
2. To understand the role and application of supervised and unsupervised learning
3. To understand the architectures of NNs building for economic applications
4. To understand the role of SOM NNs and applications in decision making
5. To learn the issues on SVM learning
|
| Course Contents |
|---|
1. Introduction to NNs and SC, mathematical model, basic learning principles.
2. Single-layer networks, perceptron – learning rule, adaptation of linear neuron.
3. Multilayer perceptrons, architectures, Backpropagation algorithms.
4. Modeling and forecasting of economic/financial time series using multilayer perceptrons.
5. Associative memories, applications to economic issues solving.
6. Recurrent NNs, RTL learning, applications to economic dynamic systems.
7. RBF NNs, architectures, learning methods.
8. NNs with unsupervised learning, competitive learning – relation ship to data mining.
9. Self organizing maps – SOM NNs, architectures, learning, applications in decision making.
10. Hybrid NNs, architecture, learning.
11. The main steps in the formulation of NNs, applications in economics and finance.
12. Machine learning, applications to data classification.
13. Regression models by support Vector Machines (SVM), application to financial high frequency time series.
14. Granular Computing (GC), principles, cloud concept, current trends in the context of probabilistic vs. intelligent (soft) computing.
|
| Recommended or Required Reading |
|---|
| Required Reading: |
|---|
SAINI, N. Review of Selection Methods in Genetic Algorithms, International Journal
of Engineering and Computer Science, 2017, vol. 6, no. 12, pp. 22261-22263.
CHARU C. Aggarwal. Neural Networks and Deep Learning. Springer International Publishing AG, 2018,ISBN 3319944622. |
MARČEK, Dušan. Pravdepodobnostné modelovanie a soft computing v ekonomike. VŠB-TU Ostrava, 2013. 300 s. ISBN 978-80-248-2955-5.
MARČEK Dušan. Supervizované a nesupervizované učení z dat: statistický a soft přístup. SAEI, vol. 45/2016, Ostrava: VŠB-TU Ostrava, 2016. 234 s. ISBN: 978-80-248-3884-7.
MARČEK, Dušan a MARČEK, Milan. Neuronové siete a ich aplikácie. Žilina: EDIS ŽU, 2006. 223 s. ISBN 80-8070-497-X.
MARČEK, D. Comparison of Predictive Statistical Learning Accuracy with
Computational Intelligence Methods. Proceedings of 15th IEEE Int. Scientific Conf.on Informatics, Eds. Steingartner et al, Poprad, pp. 254-259, 2019, ISBN 978- 1-7281-3178. |
| Recommended Reading: |
|---|
CHARU C. Aggarwal. Neural Networks and Deep Learning. Springer International Publishing AG, 2018,ISBN 3319944622.
SAINI, N. Review of Selection Methods in Genetic Algorithms, International Journal
of Engineering and Computer Science, 2017, vol. 6, no. 12, pp. 22261-22263. |
Gabor K., Kiss, A. Building Neural Networks as Dataflow Graphs. Proceedings of 2019 IEEE 15th International Scientific Conference on Informatics, Informatics 2019, pp. 216-221, ISBN: 978-1-7281-3178-8.
NIELSEN, Michael. Neural Networks and Deep Learning, Determination Press, 2015, http://neuralnetworksanddeeplearning.com/
KECMAN, V. Learning and soft computing: support vector machines, neural networks, and fuzzy logic. (Massachusetts, The MIT Press, 2001). ISBN 0-262-11255-8.
MARČEK Dušan. Supervizované a nesupervizované učení z dat: statistický a soft přístup. SAEI, vol. 45/2016, Ostrava: VŠB-TU Ostrava, 2016. 234 s. ISBN: 978-80-248-3884-7. |
| Planned learning activities and teaching methods |
|---|
| Lectures, Tutorials, Project work |
| Assesment methods and criteria |
|---|
| Task Title | Task Type | Maximum Number of Points
(Act. for Subtasks) | Minimum Number of Points for Task Passing |
|---|
| Credit and Examination | Credit and Examination | 100 (100) | 51 |
| Credit | Credit | 45 | 25 |
| Examination | Examination | 55 | 20 |