The removal of pharmaceutical actives compounds (PhACs) by nanofiltration (NF) and reverse osmosis (RO) of paramount importance in membrane separation processes. However, modeling remains a difficult approach due to the strongly nonlinear performance of the removal mechanisms of organic molecules by NF/RO. The present work features the application of neural networks based on quantitative structure–activity relationship (single neural networks “QSAR-SNN” and bootstrap aggregated neural networks “QSAR-BANN_{(Staking of 30 networks)}") for prediction of the removal of 23 pharmaceutical active compounds (PhACs). Overall, the models proposed are able to accurately correlate 599 experimental data points gathered from the literature. According to the results, the QSAR-BANN_{(Staking of 30 networks)} is a more powerful and effective computational learning machine than the QSAR-SNN. The regression coefficients “\({R}^{2}\)” and the root mean squared error “RMSE” for the QSAR-BANN_{(Staking of 30 networks)} model are estimated to be 0.9672 and 3.2810%, respectively. Moreover, QSAR-BANN_{(Staking of 30 networks)} model capabilities is showed to describe the removal of PhACS by NF/RO and its precision is compared to proposed previous models, where this comparison showed the superiority of our BANN model. The work with one class of organic compounds (PhACs) is more suitable for prediction performances NF/RO by QSAR-BANN model.

通过纳滤（NF）和反渗透（RO）去除药物活性化合物（PhAC）在膜分离过程中至关重要。但是，由于NF / RO对有机分子的去除机理具有很强的非线性性能，因此建模仍然是一个困难的方法。本工作的特点是基于基于定量结构-活性关系的神经网络（单个神经网络“ QSAR-SNN”和自举聚合神经网络“ QSAR-BANN _{（30个网络）} ”）预测23种药物的去除活性化合物（PhAC）。总体而言，所提出的模型能够准确地关联从文献中收集的599个实验数据点，并根据结果得出QSAR-BANN _{（参与30个网络）}是比QSAR-SNN更强大，更有效的计算学习机。QSAR-BANN _{（30个网络的放样）}模型的回归系数“ \（{R} ^ {2} \） ”和均方根误差“ RMSE ”分别估计为0.9672和3.2810％。此外，显示了QSAR-BANN _{（三十个网络}的样本_）模型功能来描述NF / RO去除PhACS的能力，并将其精度与建议的先前模型进行了比较，其中该比较表明了我们的BANN模型的优越性。一类有机化合物（PhAC）的工作更适合通过QSAR-BANN模型预测NF / RO。