Tree-based classification approach for dealing with complex knowledge in natural sciences

Conf.
	Proceedings of ACAI'99 - Machine Learning and Applications Chania (Greece), July 5-16, 1999
Auteurs
	D.Grosser, N.Conruyt
Abstract
	In a lot of domains based on observation, knowledge to represent and process can be very complex. For example in Systematics, the scientific discipline that studies living being diversity, descriptions of specimens are mostly structured (composite objects, taxonomic attributes), noisy (erroneous or unknown data), and polymorphous (variable or imprecise data). In this paper, we present IKBS, an Iterative Knowledge Base System for dealing with such complex descriptions. The originality of this system is to implement the scientific method in biology: experimenting (learning rules from examples) and testing (identifying new observations, improving the initial model and descriptions). This methodology is applied in the following ways in IKBS: Knowledge is acquired through a descriptive model that suits needs of experts at a semantic level. Knowledge is processed with a decision tree method in order to take into account structured knowledge introduced in the previous descriptive model of the domain. Knowledge is refined through the use of an iterative process to evaluate the robustness of the descriptive model and descriptions. The system is demonstrated on a real-world application in life science, i.e. the identification of coral specimens of the family Pocilloporidæ.