CONSTRUCTION OF FRAME HIERARCHIES USING MACHINE LEARNING

Joao Jose Furtado-Vasco, Colette Faucher, Eugene Chouraqui, Eugene Chouraqui, Eugene Chouraqui

Abstract


In this paper, we describe an architecture for helping frame hierarchy conception. This architecture is based on machine learning and cognitive psychological studies on categorizotion. Our basic assumption is that categorization should be considered as a goal-driven, context--dependent process and therefore the hierarchical organization of categories should be represented in different perspectives. The core of our architecture is a learning system of categorization that generates multi.perspective hierarchies. Concept hierarchies are, at first, generated in a probabilistic representation and after translated into a frame one.

Full Text:

PDF


DOI: http://dx.doi.org/10.7152/acro.v6i1.12660