Discrimination of Debris of Kerosene Using a Different Calculation Method of Regression Analysis

Abstract

  Sixty-one liquid kerosene samples and extracts from seventy-nine fire debris were calculated from multiple correlation coefficients (R) against two kinds of marketing kerosene as reference with a different calculation method of regression analysis by using 16 components that was eluted after n-undecane. Quantitative evaluation for debris of kerosene was performed with the R values. If the sample R value obtained was more than 0.95 using a gas chromatogram, we can judge the sample to be kerosene. For the sample R values from 0.95 to 0.85, additional information was required to judge the samples of kerosene. If the sample R value showed less than 0.85, it can be concluded that it was not kerosene.