Real-World Challenges for Data Stream Mining Workshop-Discussion

The first workshop on real-world challenges for data stream mining (RealStream 2013) will take place in Prague, the Czech Republic, on the 27th of September 2013 in conjunction with the ECML/PKDD 2013.

The objective of this workshop is to provide a forum for researchers and practitioners to discuss real-world challenges for data stream mining, identify gaps between data streams research and meaningful applications, as well as define new application-relevant research directions for data stream mining.

Workshop Format

The focus of this workshop is on presentations and discussions rather than on full written articles. The contributions range from works-in-progress, industrial experiences and a demonstration paper. The extended abstracts of the ten contributions, published as online proceedings, are available from here.

Topics of Interest

The contribution focus on real world challenges for data stream mining. Topics include:
  • Challenges and lessons learned from mining real-world data streams
    • Online data preparation and pre-processing
    • Assessing reliability of incoming data
  • Dealing with realistic data and workflows
    • End user participation to varying degrees
    • Interactive user feedback for adaptive learning
    • Reliability / correctness of feedback
    • Availability and delay of feedback
  • Integrating expert knowledge into data stream models
    • What to ask of an expert?
    • When to ask? How to set the priorities?
  • Moving from data stream algorithms towards data stream tools
    • Improving usability and trust
    • Developing autonomous, self-diagnosing data stream tools
  • Scalability of data stream mining systems