Accurate identification and quantification of organic substances represents one of the key problems during the gas chromatographic analysis. This problem is even more pronounced in cases of nontarget analyses, that is, in cases when the number of compounds is unknown. Such type of analyses is often used in environmental, pharmaceutical and food research. The number of component is unknown therefore it is difficult to optimize separation conditions. At the same time qualitative and quantitative determination is hindered due to the lack of standard materials.
Similar problems, that is, the lack of information, occur in the atmospheric chemistry as well. For instance, current three-dimensional chemistry-transport models for studying air pollution episodes in regional or global scale typically represent organic chemistry with highly simplified mechanisms. These schemes are adequate to represent near field gas phase reactivity, that is, ozone formation close to the emission sources, but fall short of providing the chemical speciation necessary for detailed understanding of the gas phase composition of the aging air parcels as well as to understand the contributions of the organic intermediates to the formation of aerosols. Major advances have been made towards the development of more detailed Master Chemical Mechanism with over 10000 reactions, however all these models are limited by the availability of the experimental reaction rates.
A quantitative structure-property relationship (QSPR) models were developed for the estimation of the missing values. A special attention was given to development of simple structural representations of chemical compounds that are easily calculated and at the same retain enough structural information of the molecule to enable development of accurate estimation models.