RSM and ANN-based Optimization of Ultrasound-assisted Extraction of Functional Components from Olive Fruit (cv Arbequina): Assessment of Antioxidant Attributes and GC-MS Profiling.

Abstract

The current study proposes an optimized ethanolic extraction for the efficient recovery of high-value components from Pakistani olives (cv. Arbequina) using response surface methodology (RSM) and artificial neural networking (ANN). The extracts were investigated for antioxidant properties and GC-MS metabolite profiling. Four factors such as time, temperature, solvent concentration, and solute weight (g/100 mL) were utilized as independent variables for determining the response (% yield). The results obtained under optimum extraction conditions such as duration (25 min), temperature (45°C), solvent concentration (65%; ethanol: water v/v), and solute (7.50 g) indicated an extract yield of 40.96% from Arbiquina olives. The analysis of variance (ANOVA) for the RSM model showed significant p-values and a correlation coefficient (R2) of 0.9960, confirming model reliability. The multilayer perceptron architecture was used in ANN, and the results were fairly consistent with the experimental findings. Arbequina olive extract (AOE) demonstrated significant antioxidant ability in terms of total phenolics, total flavonoid concentration, and DPPH radical scavenging. The GC-MS analysis of AOE revealed the presence of several bioactives, including oleic acid (36.22%), hydroxytyrosol (3.95%), tyrosol (3.32%), β-sitosterol (2.10%), squalene (1.10%), sinapic acid (0.67%), α-tocopherol (0.66%), vanillic acid (0.56%), 3,5-di-tert-butylcatechol (0.31%), and quercetin.