Development of Near Infrared Spectroscopy Models for Quantitative Prediction of the Content of Bioactive Compounds in Olive Leaves

Abstract

The objective of this work was to evaluate the ability of artificial neural networks (ANN) in near infrared (NIR) spectra calibration models to predict the total polyphenolic content, antioxidant activity, and extraction yield of the olive leaves aqueous extracts prepared with three extraction procedures (conventional extraction, microwave-assisted extraction, and microwave-ultrasound-assisted extraction). Partial least squares (PLS) models were developed from principal component analyses (PCA) scores of NIR spectra of olive leaf aqueous extracts in terms of total polyphenols concentration, antioxidant activity, and extraction yield for each extraction procedure. PLS models were used to view which PCA scores are the best suited as input for ANN based on three output variables. ANN showed very good correlation of NIRs and all tested variables, especially in the case of total polyphenolic content (TPC). Therefore, ANN can be used for the prediction of total polyphenol concentrations, antioxidant activity, and extraction yield of plant extracts based on the NIR spectra.

Keywords

• NIR spectra
• artificial neural networks
• olive leaf extracts
• conventional extraction
• microwave-assisted extraction
• microwave-ultrasound-assisted extraction