Terahertz spectra applications in identification of illicit drugs using support vector machines

Abstract

Abstract Support vector machine (SVM) was employed to classify terahertz absorption spectra for the purpose of illicit drugs identification. We successfully identified seven pure illicit drugs and found that it is a convenient and efficient method for drug identification. Then, we tried to identify drug mixtures based on the same training data, and found that the main content in a mixture can be recognized. We also tried to determine main content’s proportion of mixtures, using spectra data of various proportions, which were made for training data, based on Beer’s law. The results confirmed support vector machine to be an effective method for illicit drugs identification and content analysis.