Slides are available in the L2P system!
- Mining and retrieval in graph and network data
- Graphs are everywhere:
- social networks (facebook, google+)
- citation networks (dblp, arnetminer)
- protein structures, gene interaction networks
- program control flow, xml documents
- In this part of the lecture we discuss the following questions:
- How to measure similarity between graphs?
- How to identify groups in social networks?
- How to find frequent patterns in a graph database?
- How to classify graphs based on given observation?
- How does a real graph look like?
- How does a network evolve over time?
- Mining and retrieval in static temporal data, e.g.
- Time series and their representatios
- Similarity search on time series data
- Mining models for temporal data
- Mining and retrieval in dynamic stream data, e.g.
- Data streams occur in countless pratical applications:
- Sensor networks
- Monitoring application
- Network traffic
- Blogs, tweets, email, etc.
- Production sites
- This part of the lecture will cover
- Constraints and requirements for streaming algorithms
- General approaches for handling endless streams and evolving distributions
- Algorithms for supervised and unsupervised learning on data streams
- Questions that we will answer include
- How to track hot topics in millions of tweets?
- How to continuously update learned models?
- How to effectively handle varying data rates?
- Can we do better than using all available time? (Yes...)
Bachelor students are welcome:
Even though this course is a master course, bachelor students are welcome.
Bachelors can participate in this course if they are in the final
stage of their bachelor studies. However, it is not possible to register
for the exam (via CAMPUS or ZPA)! You will have to register separately
at our chair. After passing the exam you may hand in your results (ZPA)
as soon as you are registered as a Master student. Additionally, please send us an e-mail (containing your RWTH User-ID) to set up the access to our L2P course room (slides, announcements,...).