Application of PLS Discriminant Analysis to establish the absence of compound in electron ionization mass spectral database
T14N2
K.M. Sotnezova, A.S. Samokhin, I.A. Revelsky
Mathematical model has been developed to establish the absence of identified compound in electron ionization mass spectral library. The mathematical model (developed on the basis of PLS Discriminant Analysis) can be represented as a “black box” that provides an answer – the compound is absent in the database or not. Match factors of top ten candidates in the list of possible candidates were used as input data. More than 5000 objects (mass spectra) were used during training, validation and testing stages. Developed classification model provides correct prediction (that compound is absent in the library) in 28.4%, while only 1.2% of compounds that are represented in the database incorrectly classified as absent.
