Teaching personnel

Lecturer

• Prof. Dr.-Ing. Norbert Fuhr
• Dr. Ahmet Aker

Formalia

Targeted audience

• Angewandte Informatik Master with 6 credit points
• Komedia Master with 6 credit points
• ISE Master with 6 credit points
• Medizintechnik Master with 6 credit points

Dates

Examination Dates

Description

Information Mining deals with the extraction on implicit information from raw data (Data Mining) or text (Text Mining). The goal is the development of methods for analyzing databases and discovering useful information by means of abstraction. For this purpose, machine learning methods are applied.

Lecture material

• The major part of the course is based on the book 'Data Mining' by Ian Witten et al..
  (There is also 4th edition, but the course presents the new material in a different way).
  READ THIS BOOK!
• Other books:
  • Charu C. Aggarwal: Data Mining: The Textbook, Springer, May 2015 (PDF)
    (extensive treatment of advanced application like spatial and graph data, sequences, Web, social media, and privacy issues)
  • Jürgen Cleve, Uwe Lämmel: Data Mining. De Gruyter, 2016 (easy read, covers a subset of the Witten et al. book).
  • Shai Shalev-Shwartz, Shai Ben-David: Understanding Machine Learning: From Theory to Algorithms. Cambridge University Press, 2014.
  • Mohammed J. Zaki, Wagner Meira: Data Mining and Analysis: Fundamental Concepts and Algorithms. Cambridge University Press, 2014.

Lecture material

• Lehrangebot des Fachgebiets
• Clustering Tutorial
  Clustering Demo
• Roberto Zicari: Big Data
• Pieters: Deep Learning for NLP (Talk slides)
• Deep learning Demos:
  • Hinton's Neural Network Simulation
  • Google's Neural Network Playground
• Further reading:
  • François Chollet: On the Measure of Intelligence
  • Daniel Tunkelang: 10 Things Everyone Should Know About Machine Learning
  • SZ article on Data Analytics (in German): Das Erwachen, SZ vom 1.11.16
  • Pedro Domingos: A few useful things to know about machine learning
  • On Big Data and Data Science. Interview with James Kobielus, IBM Big Data Evangelist.
  • Susan Walsh: Between the Spreadsheets: Classifying and Fixing Dirty Data

Slides

Please note that the slides are only accessible from within the university network (use VPN)

• Chapter1-Introduction
• Chapter2-Input
• Chapter3-Output
• Chapter4-Basic Methods
• Gradient Descent
• Deep Learning 1
• Deep Learning 2
• Chapter5-Evaluation
• Chapter6-Trees and Rules
• Chapter7-Instance-based and Linear Models
• Deep Learning 3
• Deep Learning 4
• Chapter8-Data Transformations
• Chapter9-Probabilistic Methods
• Chapter11-Beyond Supervised and Unsupervised Learning
• Chapter12-Ensemble Learning
• Markov Models
• Mining Sequential Patterns
• Process Mining
• Causal Inference

Exercises

Exercise sheets can be found in the Moodle course.

The exercises will begin on 26th of October at 12:00 pm. The exercises will happen weekly through BigBlueButton (access through Moddle). You can use this link to participate. There are no solution submissions. However, instead during the exercise session you can present your solution and collect bonus points. To collect your bonus point you have to present at least 3 times successfully. A bonus point help to improve your exam results by 0.3, i.e. if you have 1.3 in your exam the bonus point make it to 1.0. Since we will be using the online system BigBlueButton for presentation it is advised you prepare your solution as power point and use PDF version while presenting. During the exercise I will read the question. You can raise your interest to present your solution. If there are more than one student interested in presenting there will be a selection through a number guess. The selected student can then upload his/her solution for presentation.