In present work, ANN (Artificial Neural Network) is compared with MRA (Multiple Regression Analysis) for their ability to predict effect of formulation parameters on product characteristics or responses. \r\nMethodology: Two factors three level full factorial design was employed for preparation of nine batches of Albendazole solid lipid Nano-particle. Lipid (Compritol ââ?¬â??ATO 888) to drug ration and amount of surfactant (Poloxamer-188) ware selected as factor X1 and X2 respectively. Besides the factorial batches, four check point batches were also prepared. Particle size and % Entrapment Efficiency (%EE) were selected as responses Y1 and Y2 respectively.\r\n ANN and MRA models were developed from result of nine factorial batches. Responses of all check point batches were predicted using both models individually. Chi-square test was performed between predicted and actual responses of both models. \r\n Response surface graphs were developed for the optimization purpose. From 3D surfaces, optimized formulation with minimum particle size and maximum %EE was selected.\r\nResult: Result data of this work conclude that, both methods are equally efficient for their prediction efficiency. However, this conclusion is not universal, ANN can be versatile model for prediction purpose. More work should be welcomed to support this result.
Loading....