Tutorials from ECML/PKDD 2007 Archive: State-of-the-Art in Data Streams and Semantic Web

The ECML PKDD 2007 tutorials archive documents the specialized sessions held in Warsaw, Poland, during the European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases. These tutorials offered in-depth explorations of core areas in machine learning and data mining, ranging from data stream mining tutorial 2007 to semantic web technologies and their integration with learning systems. Tutorials were scheduled on Monday, September 17, 2007, and Friday, September 21, 2007, surrounding the main conference program.

Data Stream Mining Tutorial (Monday, September 17, 2007)

Lecturers: João Gama (University of Porto), Mohamed Medhat Gaber (Monash University)

Summary:

• This tutorial provided an extensive overview of mining from data streams, emphasizing algorithms capable of real-time learning under memory and time constraints.
• Key themes included incremental classification, clustering evolving data, change detection, and challenges in data stream evaluation.
• Applications in sensor networks, finance, and web log analysis were highlighted.

Structure:

1. Introduction to Data Stream Models (90 minutes)
2. Algorithms for Classification and Clustering (90 minutes)
3. Change Detection and Concept Drift (90 minutes)
4. Case Studies and Applications (90 minutes)

Tutorial Notes:
Data Stream Mining Tutorial Notes (PDF)

Semantic Web and Ontology Learning (Monday, September 17, 2007)

Lecturers: Philipp Cimiano (University of Karlsruhe), Johanna Völker (University of Mannheim)

Summary:

• This session focused on techniques to automatically acquire structured knowledge from unstructured sources for use in the Semantic Web.
• Topics included ontology learning from text, semantic annotation, and the integration of statistical and symbolic methods.
• The tutorial also reviewed tools and benchmarks for evaluating ontology learning systems.

Structure:

1. Foundations of the Semantic Web (90 minutes)
2. Ontology Learning from Text (90 minutes)
3. Semantic Annotation and Knowledge Integration (90 minutes)
4. Case Studies and Evaluation Frameworks (90 minutes)

Tutorial Notes:
Semantic Web and Ontology Learning Tutorial Notes (PDF)

Probabilistic Graphical Models for Relational Data (Friday, September 21, 2007)

Lecturers: Luc De Raedt (Katholieke Universiteit Leuven), Kristian Kersting (Fraunhofer IAIS)

Summary:

• This tutorial examined the intersection of probabilistic reasoning and relational learning, introducing probabilistic graphical models suited for relational data.
• Content included Markov Logic Networks, Bayesian networks for structured domains, and inference techniques.
• Applications discussed included natural language processing, social networks, and bioinformatics.

Structure:

1. Basics of Relational Data and Probabilistic Models (90 minutes)
2. Statistical Relational Learning Approaches (90 minutes)
3. Inference and Learning Algorithms (90 minutes)
4. Applications and Future Directions (90 minutes)

Tutorial Notes:
Probabilistic Graphical Models Tutorial Notes (PDF)

Mining Complex Data: Graphs and Networks (Friday, September 21, 2007)

Lecturers: Fosca Giannotti (ISTI-CNR), Dino Pedreschi (University of Pisa)

Summary:

• The tutorial addressed algorithms and frameworks for mining complex data structures such as graphs, networks, and spatio-temporal data.
• Covered graph mining, frequent subgraph discovery, network analysis, and pattern extraction from dynamic relational data.
• Applications spanned web graphs, biological networks, and social network analysis.

Structure:

1. Introduction to Graph Data (90 minutes)
2. Graph Mining Algorithms (90 minutes)
3. Network Analysis and Dynamics (90 minutes)
4. Real-World Applications (90 minutes)

Tutorial Notes:
Mining Graphs and Networks Tutorial Notes (PDF)

Conclusion

The ECML PKDD 2007 tutorials archive captures a turning point in the teaching of state-of-the-art methods, bridging foundational learning techniques with the emerging demands of complex, dynamic, and semantic-rich data. These tutorials remain valuable as reference points for understanding the evolution of data stream mining, semantic web technologies, relational graphical models, and network analysis. For those interested in how these foundational tutorials have evolved into today’s state-of-the-art AI, see our modern AI tutorials collection, which explores advances in deep learning, graph neural networks, and large-scale knowledge representation.