Abstract

Introduction Despite enormous efforts in quantification, our understanding of many of the Earth's systems remains non-axiomatic; the systems are 'open' and consequently it is not possible to deduce all outcomes from known laws. Science must therefore adopt a manner that encourages the creation or uncovering of new knowledge (Baker, 1999; Takatsuka & Gahegan, 2001). For this reason alone and completely uncoupled from concerns about increasing data volume: '.'t is vital that knowledge discovery methods can be brought successfully to bear across the geosciences. The focus of this position paper is on the relationship of data mining and knowledge discovery to the different approaches used for scientific inference. It is via an understanding of this relationship that we can categorize the kinds of knowledge that can be discovered or learned, and gain an understanding of the roles played in the knowledge discovery process by the domain expert and the computational tools used. Many differe

Citations