Three-way data modelling advantages to solve classification issues

S.M. AZACARATE; A.A. GOMEZ; J.M. CAMIÑA; H.C. GOICOCHEA

Halifax

Congreso; XVII Chemometrics in Analytical Chemistry Conference - CAC 2018; 2018

Chemometrics in Analytical Chemistry

Over the last years, in multivariateclassification setting, second-order classification methods have been appliedas a need to solve a high complexity data structure [1]. Notwithstanding, theirapplications has not been motivated on the need to increase the data order as astrategy to solve a classification problem [2]. In this work, the possibilityto increase order data from two to three ways to achieve a potentialclassification when a first-order classification is not possible isdemonstrated. For that, first-and second-order fluorescence spectroscopy datasets -simulated and experimental- were acquired. Simulated two- and three-waydata involving three target classes were generated. Later, experimental datawas acquired from 80 mayonnaise samples to evaluate spoiled mayonnaiseaccording to different five-storage time. On the one hand, samples were excitedto 320 nm and the spectra were recorded from 300 to 500 nm. After, fluorescenceexcitation spectra were recorded between 230 and 400 nm, and emissionwavelengths from 300 to 500 nm. PLS-DA and N-PLS-DA were employed for first andsecond-order data, respectively. Later, model predictive ability were evaluatedthrough some figures of merit, e.g. sensitivity, specificity, precision andaccuracy. The results showed a poor classification for first- order data with ahigh error rate of 78 %, which was awesomely improved when second-order datawere applied. Certainly, it was possible to attain a significant improvement onresults and a positive impact on analytical figures of merit when second-orderdata were analyzed. Thus, it can be demonstrated that the way in which the dataare generated has a significant effect on the classification. References [1] J.M. Amigo, F. Marini, MultiwayMethods, in: F. Marini (Ed.), Data Handling in Science and Technology,E-Publishing Inc., Amsterdam, 2013, pp. 283?309. [2] E. Salvatore, M.Bevilacqua, R. Bro, F. Marini, M. Cocchi, Compr. Anal. Chem. 60 (2013) 339-379. Acknowledgement:Authors are grateful to UNL, UNLPam, CONICET and ANPCyT.