Optimization of Bioactive Compound’s Extraction Conditions from Beetroot by Means of Artificial Neural Networks (ANN)

Authors

  • Raquel P.F. Guiné CI&DETS/CERNAS Research Centre, Dep. Food Industry, Polytechnic Institute of Viseu http://orcid.org/0000-0003-0595-6805
  • Mateus Mendes Polytechnic Institute of Coimbra – ESTGOH / Institute of Systems and Robotics of the University of Coimbra, Dep. of Electrical and Computer Engineering, University of Coimbra, Portugal
  • Fernando Gonçalves CI&DETS/CERNAS Research Centre, Dep. Food Industry, Polytechnic Institute of Viseu

Keywords:

Phenolic compounds, Antioxidant activity

Abstract

The present work used Artificial Neural Network (ANN) models to correlate beetroot extraction conditions with total phenolic compounds (TPC), anthocyanins (ANT) and antioxidant activity (AOA). The input variables were extraction time, type of solvent, solvent volume/sample mass (VMR) and order of extraction. The ANN models produced showed very good accuracy (R > 94 %), being suitable for data mining using weight analysis techniques. The experiments involved variable conditions: solvents (Methanol, ethanol:water and acetone:water), extraction times (15 and 60 min), VMR (5, 10 and 20), order of extract (3 sequential extractions). The TPC were evaluated by the Folin-Ciocalteu method, ANT by the SO2 bleaching method and AOA following the ABTS method. The experimental results showed that the extracting solutions used in this study exhibited similar extraction efficiency for TPC, but not for AOA. Also, the results allowed concluding that a higher VMR originated extracts with higher amounts of TPC and AOA.

Downloads

Published

2019-12-16

Issue

Vol. 21 No. 4 (2019): CIGR Journal

Section

VI-Postharvest Technology and Process Engineering