The non-linear relations exist between the temperature,pressure,and concentration with the flux and retention coefficient of Agaricus Blazei Murill polysaccharide solution during ultrafiltration(UF) separation.The models of UF were built by back-propagation algorithm based on Neural Network Toolbox(NNT)of MATLAB.The parameters of model were discussed,such as the number of hidden layer node,learning method and learning rate.The results show that the network system was 3×13×1,and there is a good agreement between experimental data and modeled data with prediction precision 1.67% and 3.75%.
以多氯联苯分子的范德华体积、偶极矩、平均分子极化率和分子椭圆度等分子结构信息为基础,通过线性多元回归建立了它们与正辛醇—空气分配系数K OA的定量结构—性质关系(QSPRs),预测了一些多氯联苯的K OA。同时研究了分子结构数据对K OA的影响,结果表明对K OA影响最大的是范德华体积和平均分子极化率。苯环上氯原子数越多,对-log K OA的影响越大,非邻位氯取代的多氯联苯分子的-log K OA较大,其次是单邻位取代的多氯联苯分子的-log K OA,双邻位取代的多氯联苯分子的-log K OA较小。
