X-ray Powder Diffraction and Supervised Self-Ogranizing Maps as Tools for Forensic Classification of Soils.

Abstract

Due to its transferability, the soil has been commonly used as evidence in criminal investigations. In this work, 172 soil samples taken from five urban parks from the town of Tetovo (North Macedonia) and from additional four rural locations in its vicinity. The soil samples were examined using X-ray powder diffraction. The collected diffractograms were used for development of classification models based on supervised self-organizing maps. The examination generalization performances of the developed models showed that they were able to correctly classify between 95.6 and 97.8% of the samples from the independent test set. The influence of the weather and the seasonal changes on the composition of the soil was also examined. For this purpose, three years after the initial soil samples were collected, additional 28 samples were analyzed from different location. The best models presented in this work were able to successfully classify 27 of these additional samples.