Prediction of mole fraction solubility of drugs in supercritical carbon dioxide by using artificial neural network (ANN) method in comparison with semi-empirical models

In this study, the solubility of drugs in supercritical carbon dioxide (scCO2) was represented by using the artificial neural network (ANN). Prediction results were compared with nine various models. This research has used a multilayer feedforward network with Levenberg Marquardt backpropagation training for prediction. Original data were divided into two parts where 70% of data was used as training data and remaining 30% of data was used for testing. In this method, inputs are pressure and temperature. The number of neurons is set at four. Using ANN for predicting mole fraction solubility of drugs in scCO2 leads to less deviation compared to other proposed models and it is in good agreement with experimental data

کلیدواژه ها:

Drug solubility ، Neural network ، Prediction ، Supercritical carbon dioxide