Artificial Neural Network Modeling of Ultrasound-Assisted Polysaccharides Extraction from Potentilla anserina and Anti-Platelet Aggregation Activity

Abstract

In this study, Ultrasound-assisted extraction (UAE) was used for polysaccharides extraction from Potentilla anserina. A computer-stimulated artificial neural network (ANN) was developed to get a correlation between the input variables and the output parameter. Finally, the optimal process conditions were obtained as follows: extraction temperature 55 °C, extraction time 55 min, ratio of liquid to solid 20:1, power 175 W. Under optimized conditions, ultrasound-assisted extraction had obviously higher yield of polysaccharides than the traditional heat reflux method. The optimization procedure showed a close interaction between the experimental and simulated values for polysaccharides extraction. The R2 (0.99271) and MSE (0.0425) values of model suggested good fitness and generalization of the ANN. Moreover, the results also indicated that polysaccharides have inhibitory effect on ADP-induced platelet aggregation.