"Machine Learning and Knowledge Extraction Techniques for Natural Resources Inventory"
A-Xing Zhu, Department of Geography, University of Wisconsin at Madison

This workshop introduces the techniques for extracting knowledge on relationships between natural resources and their environmental conditions. The extracted knowledge is needed for conducting predictive mapping of natural resources (such as soils) and for mapping the susceptibility of natural hazards (such as landslides). The techniques to be discussed include: neural networks, case-based reasoning, personal construct-based interview techniques, decision-trees, and noise-reduction techniques for spatial data mining. Each of the techniques will be introduced and discussed using a real application (soil mapping). Attendants will also gain a hand-on experience of using some of the techniques. Software and real world data set will be provided.

• PDF file of paper presented