Fuzziness vs. probability in a data mining application for soil classification

Data mining methods have been proven effective in extracting knowledge from existing data sources for the classification of soils. Previous studies have suggested that soils are spatial entities with fuzzy boundaries and prompted the development of data mining methods to extract knowledge that allows for fuzzy classifications of soils. This paper first looks at the nature of soil classification from the perspective of cognitive psychology. It then examines data mining methods used for fuzzy soil classification. It notes that some of the methods are inherently hybrids that combine statistical measures with fuzzy models on sound cognitive bases. This paper reflects upon the long lasting debate on fuzziness versus probability for modeling uncertainties and suggests that hybrid models are valid both practically and cognitively. At last, some preliminary results are reported in comparing pure probabilistic methods (Bayesian), a fuzzy method, and two hybrid approaches to knowledge discovery for soil classification that supports the suggestion.