Přístupnostní navigace
E-application
Search Search Close
Course detail
FEKT-BPC-UIMAcad. year: 2024/2025
The course is oriented on commonly used methods in the field of artificial intelligence: artificial neural networks, fuzzy logic and fuzzy inference systems, cluster analysis. Both theoretical (basic principles of each method) and practical (applications to the problem of classification, regression and clustering) aspects are discussed. The theory is discussed in direct connection with practical examples. All computational techniques are practiced using the Python environment. The course prepares students to independently use the given methods for data analysis in their own scientific or routine work.
Language of instruction
Number of ECTS credits
Mode of study
Guarantor
Department
Entry knowledge
The knowledge on the Bachelor´s degree level is requested, namely on numerical mathematics. The laboratory work is expected knowledge of Python.
Rules for evaluation and completion of the course
The conditions for successful completion of the course are specified in the annually updated decree of the course guarantor.
1) Team project (max. 25 points):
- Preparation of an original team project solution and its defence at the end of the semester (according to the guidelines).
- the completion of the assignment and the quality of the presentation of the results by all team members will be evaluated
- plagiarism will result in 0 credit
2) Final exam (max. 75 points):
- Combined form (written and oral)
- three parts in total, each for a maximum of 25 points
Conditions for credit and admission to the final examination:
- obtaining a non-zero number of points for the team project
- a maximum of two excused absences
Conditions for successful completion of the course:
- obtaining credit
- obtaining at least 36 points in the exam
- obtaining a total (i.e. team project and exam) of at least 50 points
Aims
The goal of the course is to provide the students with sufficient knowledge from artificial intelligence area and to present them the possible use of modern tools of artificial intelligence in acquisition, processing and analysis of biomedical data.Candidates will get knowledge and skills in area of artificial intelligence applications. He will be competent to apply some widespread methods for real tasks solving, naimly to process and analyse data. During written examination, it is verified, whether the student is able to:- discuss basic terms from artificial intelligence area,- describe basic methods in this area,- discuss advantages and disadvantages of particular methods,- select and apply appropriate tools to solve the task,- estimate the quality of obtained result and present it in a proper way,- interpret obtained results.
Study aids
Prerequisites and corequisites
Basic literature
Recommended reading
Elearning
Classification of course in study plans
Lecture
Teacher / Lecturer
Syllabus
1. Introduction to artificial intelligence. Areas of application: classification (into two or more classes), regression and clustering. Overview of machine learning algorithms.
2. Preparation of measured data: feature-based description, normalization, selection of informative features, feature decorrelation.
3. Quality assessment of classification, regression and clustering results.
4. Artificial neural networks, perceptron and its characteristics. Neuron as a classifier. Linear vs. non-linear task.
5. Learning a neuron with binary and real inputs and outputs, single layer perceptron.
6. Multilayer feedforward network, error back propagation algorithm.
7. Cluster analysis, hierarchical cluster analysis methods.
8. Non-hierarchical cluster analysis methods, k-means algorithm, fuzzy c-means algorithm.
9. Fuzzy sets, fuzzy relations, fuzzy logic.
10. Approximate inference. Fuzzy inference systems.
11. Examples of using artificial neural networks, clustering and fuzzy inference systems to solve real-world problems.
Exercise in computer lab
1. Basics of vectorization and matrix operations
2. Hierarchical data clustering
3. Non-hierarchical data clustering
4. Fuzzy data clustering
5. Feature reduction and principal component analysis
6. Perceptron design (without learning)
7. Neural network design (without learning)
8. Delta rule
9. Forward network learning
10. Feature reduction, model validation and evaluation of classification results
11. Fuzzy inference