DTAI Projects
FWO: Graph logic: Representation, Inference and Learning
Period: 01-2007 → 12-2010
Subgroup: ml
Type: project
Members:
- Hendrik Blockeel
- Jan Ramon
Relational data mining concerns the analysis of data where individual objects are not independent but related to each other. Among methods for relational data mining we can distinguish methods based on first order predicate logic, methods based on the relational database model, and methods based on graphs. Among these, the logic-based methods have the largest expressiveness with respect to the background domain knowledge that can be exploited, while graph-based methods focus more on the graph structure of the relationships. This project aims at developing a "graph logic" for data mining applications in mind. Such a graph logic should allow us to develop data mining methods that combine the strengths of graph-based and logic-based methods.
← return to the projects
