CHEMOMETRIC TOOLS USED IN THE PROCESS OF FIRE INVESTIGATION

Fire debris analysis is one of the most challenging steps in fire investigation. Data analysis is subjected to human interpretation by comparing with a reference database. In recent years, chemometric tools have been successfully applied to data in order to avoid misunderstandings and make data interpretation less subjective. In the present study, two ignitable liquids (ILs) (diesel and ethanol) as well as two substrates (cork from Quercus suber bark and cotton sheet) were used. Fire debris samples were prepared following the modified procedure of Destructive Distillation Method for Burning. Sampling was performed in delayed times. In accordance with a progressive approach to data, multivariate statistical analyses, such as unsupervised hierarchical cluster analysis (HCA) and principal component analysis (PCA) as well as supervised linear discriminant analysis (LDA) were applied. The aim of the current research is to investigate whether HS-MS eNose is able to detect different ILs among fire debris samples containing different substrates after delayed times of sampling.