Modeling of Drug Release from Matrix Tablets with Process Variables of Microwave-Assisted Modification of Arrowroot Starch Using Artificial Neural Network

Abstract

The objective of this study was to model the drug release property in terms of process variables of microwave-assisted modification of arrowroot starch using artificial neural network (ANN). The water content, microwave power and heating time were used as process variables for modification of arrowroot starch and the mean dissolution time was used as dependent variable. The correlation between process variables and dependent variable was examined using feed-forward back-propagation neural networks. The ANN model was optimized by considering goodness-of-fit and crossvalidated predictability. A “leave-one-out” cross-validation revealed that the neural network model could predict MDT values from matrix tablets with a reasonable accuracy (predictive r2 of 0.824 and predictive root mean square error of 19.53). The predictive ability of these models was validated by a set of 4 formulations that were not included in the training set. The predicted and observed MDT were well correlated.