Prediction of in vitro release of nanoencapsulated phenolic compounds using Artificial Neural Networks

Abstract

In Vitro Release modeling (IVR) of nanoencapsulated phenolic compounds (PC) is complex, due to the number of factors involved in the process. Artificial Neural Networks (ANN) are useful tools for its prediction because they consider the effect of all factors on the response. The release at 5h is crucial in kinetics because, in most cases, it is an equilibrium point leading to a constant phase. The objective of this investigation was to predict the IVR of nanoencapsulated PC at 5h using ANN. A database with information from the scientific literature was used. This model permits mathematical correlation of the IVR at 5h with eleven factors. The optimal network configuration consisted of one hidden layer with one neuron. A mathematical model was obtained with a Mean Square Error (MSE) of 0.0516 and a correlation coefficient (r) of 0.8413.

Keywords

• phenolic compounds
• ultrasound
• nanoencapsulation
• Artificial Neural Networks (ANN)