genetic algorithm using dd-simca one class through ft-ir spectroscopy to classification of listeria samples

Abstract: nutrition-caused listeriosis is one of the most serious and most severe nutrition-transferable diseases that can be considered a threat to human health. owing to the high resistance of listeria monocytogenes against environmental conditions, existence of this bacterium in raw or processed foods is also possible and the appropriate evaluation and monitoring is necessary actions and regulations from the beginning of the production cycle to the consumption of the nutrients. the diagnosis of pathogens in food safety is essential [1,2].this study investigated the classification of listeria samples using ir spectroscopy. the mid-ftir spectra were pretreated by baseline corrected, multiplicative scatter correction (msc) transformation to eliminate the baseline shift deletion and multiplicative effect of scattering and orthogonal signal correction (osc) which removes unrelated or orthogonal systematic variation from the spectral data. in the next step, from 792 wavelengths, 202 wavelengths were selected by the genetic algorithm (ga) algorithm as a feature selection procedure for dd-simca. the dataset of 1717 samples was split into two subsets as calibration and validation sets through randomly. table 1 (graphical abstract) is shown summary of dataset after outlier detection, ga variable selection and msc preprocessing. the results are summarized in table 2 (graphicalabstract). therefore, the aim of this research is to introduce an easy, fast, and low cost method to identify listeria based on spectroscopy studies and using chemometrics method. table 2- summary of the basic features of the models.