ABSTRACT
Several factors such as moisture state, particle geometry, gradation parameters, fines content, and the nature of stress paths contribute to the directional dependency of material properties in the granular layers. The motivation for this study was to evaluate the applicability of pattern classification techniques to provide class discriminatory information of the laboratory observations. For this purpose, several experimental permutations were subjected to variable dynamic confining pressure stress path tests to study the synergistic influence of different factors on the anisotropic behavior of aggregate systems. The laboratory tests and the post processed data were in turn used to evaluate the relevance of pattern classification techniques to unravel physically meaningful information of the multi-dimensional dataset. The results of this effort will be instrumental for the practitioners to potentially reduce the number of features needed to be determined in the laboratory for a refined and cost-effective testing protocol.
