Short Course: Selected Topics in Data Science, 2 ECTS, June 4-8
Learning Outcomes: The student learns the state-of-the-art data analysis methods. He/she also knows how to apply these techniques to deal with the practical problems in the areas of system modeling, data anomaly detection, time series prediction, data clustering, optimization, pattern recognition, etc.
Contents: The typical data mining, parameter estimation, and optimization algorithms including least square method, k-means clustering technique, artificial immune systems, genetic algorithms, etc. are introduced. Their applications in solving a large variety of engineering problems will be demonstrated as well.
Assessment Methods and Criteria: Examination and computer exercises are graded.
Study Materials: Lecture notes and selected papers from journals and conference proceedings.
1. Least Square Method for Parameter Estimation
Theory of least square method, curve fitting, and polynomial-based approximation. Applications of least square method in parameter estimation and system identification.
2. Artificial Immune Systems in Data Anomaly Detection
Principles of artificial immune systems and negative selection algorithm. Anomaly detection in time series data and fault diagnosis. Clonal selection algorithm. Artificial immune networks in data clustering.
3. Data Clustering Methods and Applications
Basic concepts of data clustering. Distances in data similarity assessment. K-means. Fuzzy k-means. Mountain clustering method. Subtractive clustering method.
4. Introduction to Optimization Theory, Methods, and Applications
Essential concepts and theory of optimization. Convex optimization methods. Evolutionary computation. Swarm intelligence methods, Hybrid optimization algorithms. Multi-objective optimization. Constrained optimization. Applications of optimization techniques in data mining.
Course Schedule: Lectures: 10:00-12:00 and computer exercises: 14:00-16:00 from Monday to Friday (June 4 to 8, 2018).
Registration will open on Monday May 21