DTAI Projects
ESPRIT project ALADIN: Applied Logic for Advanced Data mining In iNdustry
Period: 10-1998 → 09-2000
Subgroup: ml
Type: project
Members:
- Hendrik Blockeel
- Maurice Bruynooghe
- Luc Dehaspe
- Luc De Raedt
The ALADIN project aims at the dissemination of ILP (Inductive Logic Programming) in the data mining community. It is a medium sized project with a duration of 24 months. Modelling has a key role in data mining, but current modelling techniques (rule induction, decision tree induction, association rule induction) have two important shortcomings:
1. they are unable to integrate explicit formulations of background knowledge, and
2. they are unable to work with multiple tables.
ILP is a new family of modelling techniques, which overcomes these problems, and can therefore improve the results of data mining and extend its applicability to new domains. ILP offers a knowledge rich approach to modelling: the data miner can incorporate domain knowledge into the modelling process in the form of a logic program which spells out the links and relationships already known to be in the domain of application, i.e. the background knowledge.
One of the goals of this project is to obtain a better understanding of when to use which ILP technique. To this end, the project will compare ILP rule induction (Progol), with ILP decision tree induction (Tilde), and ILP association rule discovery (Claudien).
ILP is one of the very few areas within artificial intelligence where European research is taking the lead. The best way to turn this scientific advance into a European competitive advantage is to incorporate ILP techniques into the major European data mining tool Clementine. In addition, two application partners will investigate different vertical applications and will in this way guarantee the applicability of the inductive logic programming techniques.
← return to the projects
