Extracting Generic Cooking Adaptation Knowledge for the TAAABLE Case-Based Reasoning System

Emmanuelle Gaillard (1) , Emmanuel Nauer (1) , Marie Lefevre (2) , Amélie Cordier (2)

1 ORPAILLEUR - Knowledge representation, reasonning
2 SILEX - Supporting Interaction and Learning by Experience

This paper addresses the issue of interactive adaptation knowledge acquisition. It shows how the expert's involvement in this process can improve the quality and usefulness of the results. The approach is defended in the context of \taaable, a \cbr system which adapts recipes to user needs. In \taaable, adaptation knowledge takes the form of substitutions. A datamining process allows the discovery of specific substitutions in recipes. A second process, that must be driven by an expert, is needed to generalise these substitutions to make them usable on other recipes. For that, we defend an approach based on closed itemsets (CIs) for extracting generic substitutions starting from specific ones. We focus on a restrictive selection of objects, on a specific filtering on the form of the CIs and on a specific ranking on support and stability of the CIs. Experimentations demonstrate the feasibility of our approach and show some first results.

Mots clés: adaptation knowledge discovery; interactive knowledge acquisition; closed itemset; cooking

Domaines: Intelligence artificielle [cs.AI]

Emmanuelle Gaillard, Emmanuel Nauer, Marie Lefevre, Amélie Cordier. Extracting Generic Cooking Adaptation Knowledge for the TAAABLE Case-Based Reasoning System. Cooking with Computers workshop @ ECAI 2012, Aug 2012, Montpellier, France. ⟨hal-00720481⟩