基于支持向量机的药物诱导磷脂质病预测模型  

A prediction model of drug-induced phospholipidosis based on support vector machine

作　　者：解扬[1] 张会[1] 杨胜勇[1] 

机构地区：[1]四川大学华西医院生物治疗国家重点实验室,四川成都610041

出　　处：《化学研究与应用》2011年第6期696-701,共6页Chemical Research and Application

基　　金：国家自然科学基金(20872100)资助

摘　　要：本文应用一种组合遗传算法和共轭梯度法的支持向量机(GA-CG-SVM)方法建立了药物诱导磷脂质病分类预测模型。首先对描述符进行了优化,选出了19个描述符用于模型的构建,所建模型对训练集的预测准确率为81.6%,对测试集的预测精度为87.5%,说明所建SVM分类模型不仅能正确预测训练集药物诱导的磷脂质病,也对其他化合物具有很好的预测能力。In this study,a support vector machine（SVM）method combined with genetic algorithm（GA）and conjugate gradient（CG）algorithm was used to build a classification prediction model of drug-induced phospholipidosis.The descriptors were optimized firstly and 19 descriptors were finally selected to construct the SVM model.The overall prediction accuracy of the established model for the training set is 81.6%,and that for the test set is 87.5%.These show that the SVM model not only can correctly predict the training set drug molecules but also has good prediction ability to compounds outside of the training set.

关 键 词：支持向量机 遗传算法 共轭梯度法 磷脂质病 

分 类 号：R991[医药卫生—毒理学]