Data Mining and Probabilistic Reasoning

Lecturer

Dr. Gjergji Kasneci

Abstract

Data arising from business transactions, scientific measurements and other forms of content-creation calls for automatic data mining and pattern recognition techniques that allow us to efficiently make sense of this data. At the same time these techniques should be able to handle uncertainty, as data from measurements may be imprecise and user-generated content may be unreliable.

This lecture will introduce the main concepts of data mining and probabilistic reasoning, ranging from basic probability and information theory to popular classification and clustering algorithms. An introduction to the exciting area of graphical models and probabilistic inference will highlight the link between uncertainty and probabilistic learning models.

Topics

raphical models

Literature

I. H. Witten, E. Frank, M. A. Hall: Data Mining - Practical Machine Learning Tools and Techniques (Chapters 1 - 6)

C. Bishop: Pattern Recognition and Machine Learning (Chapters 1 - 4, 8, 9)

T. M. Mitchell: Machine Learning (Chapters 3 - 6, 8, 10)

P. Flach: Machine Learning – The Art and Science of Algorithms that make Sense of Data (Chapters 1 – 3, 5 – 11)

D. J. C. MacKay: Information Theory, Inference and Learning Algorithms (Chapters 1 - 6)

Tools

The Weka Toolkit

Timetable

• Lectures:
  • Tuesdays 13:30-15:00 in Room H-E.51
  • Every second Thursday 11:00-12:30 in Room  H-2.57
• Exercises: 
  • Every second Thursday 11:00-12:30 in Room  H-2.57
  • The assignments are available in the "Materialien" folder in the "Interner Bereich"

We thank all students for their help in pointing out errors in the slides!

The exam timetable and room information are now available in the "Interner Bereich"

Exam

Condition for exam admission: oral presentation of at least two solutions during the tutorials 

Form of exam: oral exam at the end of the term