Abstract

The present research studies the geographical classification of olive mill wastewater (OMWW) samples using FT-MIR spectroscopy and chemometric recognition algorithms. The study focuses on samples collected from two regions in Morocco: BeniMellal-Khenifra and Fes-Meknes. Principal Component Analysis (PCA) has been applied to the spectral data, revealing distinct clustering of OMWW samples based on their geographical origin. Additionally, Partial Least Squares Discriminant Analysis (PLS-DA) has been employed to develop a classification model for accurate sample categorization. The PLS-DA model presented high sensitivity and specificity, achieving perfect classification rates during the calibration phase. Validation results show the model's capability to accurately identify the majority of samples, with a minor misclassification. These results confirm the potential of FTMIR spectroscopy and chemometric recognition algorithms for authentication, and traceability in the olive oil industry. Further research can explore the extension of this methodology to other regions and agricultural by-products, incorporating advanced machine learning algorithms for enhanced classification accuracy. Overall, the present study contributes to enhance environmental waste management research by addressing OMWW disposal concerns and promoting sustainable practices on the valuation possibilities.

Keywords: Olive Mill Wastewater, OMWW, FTMIR, data analysis, traceability, classification and discrimination

Article language: En