论文部分内容阅读
为了预测药物透血脑屏障的活性,计算表征分子组成和拓扑等特征的87个分子描述符,经遗传算法筛选,参与建立基于支持向量学习机(SVM)的药物透血脑屏障活性分类模型。在模型训练中用网格搜索法确定核函数的两个重要参数C和γ,同时用5重交叉验证模型,结果证明模型预测能力较高,交叉验证的预测正确率达85.6%。
In order to predict the activity of drug through the blood-brain barrier, 87 molecular descriptors, which characterize molecular composition and topological characteristics, were calculated and screened by genetic algorithm to establish a drug-permeable brain barrier active classification model based on Support Vector Machines (SVM). In the model training, grid search was used to determine two important parameters of kernel function C and γ. At the same time, the model was validated by 5-fold cross-validation. The results showed that the model had higher predictive ability and the correct rate of cross-validation was 85.6%.