Yazar "Genis, Huseyin Efe" seçeneğine göre listele

Listeleniyor 1 - 1 / 1
  • [ X ]
    Öğe
    Rapid discrimination of Turkish commercial hazelnut (Corylus avellana L.) varieties using Near-Infrared Spectroscopy and chemometrics
    (Elsevier, 2022) Ayvaz, Huseyin; Temizkan, Riza; Genis, Huseyin Efe; Mortas, Mustafa; Genis, Duygu Ozer; Dogan, Muhammed Ali; Nazlim, Burak Alptug
    Classification of hazelnuts as per their varieties is traditionally made based on subjective phenotypic observations. However, various factors easily affect these observations, including environmental conditions and the growth phase of hazelnuts, making it difficult to separate the varieties. Furthermore, the distinction among the varieties becomes far more challenging in the case of shelled hazelnut kernels, where trade is much more common than hazelnuts in-shell. Nevertheless, literature studies imply that there are apparent differences among the chemical compositions of different hazelnut varieties along with the visual appearances. Thus, developing an analytical method allowing these differences to be determined quickly and objectively would be valuable. Accordingly, the potential of using Near-Infrared Spectroscopy (NIRS) with chemometrics as a fast, convenient, and reliable method in discriminating commercially important hazelnut varieties grown in Turkey has been investigated in this research. A total of 280 hazelnut samples belonging to seven distinct varieties in Turkey were employed in the study. NIR spectra of the hazelnuts were collected from three physical states of hazelnuts as (1) in-shell, (2) shelled, and (3) white interior of split hazelnut kernels. Using the collected spectra and chemometric methods (Partial Least Squares-Discriminant Analysis (PLS-DA) and Logistic Models), calibration models were developed using a training set (n = 210) and validated with a test set (n = 70). The results have indicated that all NIR spectra-based PLS-DA and Logistic Models with MultiClass Fisher's Linear Discriminant Analysis (FLDA) pre-processing technique could accurately determine the variety of unknown hazelnuts at 100 % in either of the three tested physical states of the samples. Accordingly, these approaches can be utilized by the sector that is concerned about economic loss as more reliable and objective methods compared to the traditional analysis of hazelnut varieties, which are generally made via subjective visual assessment.