All programmes > Course: Data Mining

Data Mining (B-KUL-H02C6A)

4.0 ECTS English 36.5 Second termSecond term Advanced
Davis Jesse
POC Artificial Intelligence

Aims

In many branches of science and industry today institutions gather more
data than they can digest. This simple fact is the driving force behind a hybrid research field called data mining or knowledge discovery in databases (KDD). Data mining combines techniques from statistics, databases, pattern recognition, computer graphics, and artificial intelligence to increase the digestive capacities of more traditional data analysis tools. The course covers the different steps in the essentially cyclic and interactive data mining process. While the full data mining process provides a framework throughout the course, modeling techniques, which are at the heart of this process, receive most attention.

Previous knowledge

Bachelor or Master level with at least basic knowledge of computers, algorithms and data structures.
Knowledge of Machine Learning techniques.

Course material

Articles and literature
Slides, transparencies, courseware
Multimedia

Is also included in other courses

Activities

0.8 ects. Data Mining: Practical Sessions (B-KUL-H00I0a)

0.8 ECTS English 20.0 Second termSecond term
Davis Jesse
POC Artificial Intelligence

Aims

Hands-on experience is provided with a case study involving professional data mining software.

3.2 ects. Data Mining: Lecture (B-KUL-H02C6a)

3.2 ECTS English 16.5 Second termSecond term
Davis Jesse
POC Artificial Intelligence

Content

Session 1,2: Introduction (M. Van Hulle)
Overview of the KDD process
Data mining objectives, primary tasks
Data selection, preprocessing, transformation
Building blocks of data mining
Data mining tool classification
Session 3,4: Data Transformation (M. Van Hulle)
Linear-and non-linear projection techniques
PCA, ICA, MDS, self-organizing topographic maps, GTM
Session 5: Feature selection & feature extraction (M. Van Hulle)
Unsupervised & supervised techniques
Data inequality theorem
Filters, wrappers
Maximum relevance minimum redundancy feature selection
Case study
Session 6-9: Frequent pattern discovery (L. Dehaspe)
Representations and tasks
Algorithms Empirical results
Generic framework
Case study
Session 10,11: Graph mining (L. Dehaspe)
Structure and properties of large graphs
Fitting network models
Case study
Session 11,12: Data stream mining (L. Dehaspe)
Paradigms for knowledge discovery from evolving data
Evolution in association rules
Change mining
Session 13: Visual mining (M. Van Hulle)
Case studies

Evaluation

Evaluation : Data Mining (B-KUL-H22C6a)

Mode of evaluation : Written
Category : final examination during examination period

Explanation

Closed book + case study evaluation.