Heroin samples [76 South East Asian (SEA), 20 non-South East Asian (non-SEA) and 92 of uncertain origin] were analysed by inductively coupled plasma mass spectrometry for 73 elements. Hierarchical cluster analysis (HCA) grouped together the SEA and many of the non-SEA sub-groups, and grouped together elements with similar chemical properties. K-means cluster, correlation and principal component analyses supported the HCA. Logistic models were created using continuous, discrete and a combination of continuous and discrete elemental data in order to predict the SEA or non-SEA origin of samples. The predictive values of the models were 68–100%, the most successful models simultaneously using both continuous and discrete concentration data.
No takes yet. Share an insight, caveat, or question.
Myors et al. (1998) studied this question.
Synapse has enriched 5 related papers. Consider them for comparative context: