人工神经网络预测盐酸帕罗西汀缓释微丸的药物释放

Prediction of the Drug Release from Sustained Release Pellets of Paroxetine Hydrochloride by an Artificial Neural Network

目的利用人工神经网络对盐酸帕罗西汀缓释微丸的释药行为进行预测。方法设计20个处方,其中16个处方作为训练处方,其余4个处方作为测试处方,制备盐酸帕罗西汀膜控释微丸,进行释放度检查。以致孔剂PVPK30的用量、包衣增重作为自变量,考察药物在各个取样点的累积释放量作为输出,建立盐酸帕罗西汀缓释微丸释药行为的人工神经网络预测模型。通过线性回归法、相似因子法、AIC法评价人工神经网络的预测能力。结果通过实测数据和BP神经网络预测结果比较,验证了人工神经网络的预测精度达0.989 9。结论用人工神经网络对盐酸帕罗西汀缓释微丸的释药行为进行预测,拟合度较高,从而为盐酸帕罗西汀缓释微丸的处方优化和释药行为预测提供了可行的依据。

OBJECTIVE To use an artificial neural network （ANN）to predict drug release from sustained release pellets of paroxetine hydrochloride. METHODS As model formulations,20 formulations of paroxetine pellets were prepared, 16 of which were set as training data,while the other 5 were for prediction. The amount of PVPK3o and coating level were selected as casual factors,and the accumulative drug release in each sampling time was used as response variables. A set of release parameters and causal factors were used as tutorial data for ANN and analyzed by computer. The predictive ability of the ANN model was assessed by comparing the linear regression equations ,similarity factors （f2） and AIC values of predicted against observed property values. RESULTS Comparing the predicted values with experimental data, R2 of the linear regression was 0. 989 9. CONCLUSION The prediction technique incorporating ANN showed a fairly good agreement between the observed values of release parameters and the predict results. Therefore, ANN provides a feasible way of optimizing and estimating drug release from sustained release pellets of paroxetine hydrochloride.