论文部分内容阅读
将Vapnik提出的支持向量机(support vector machine,简称SVM)算法用于化合物活性的模式识别研究。SVM算法是特别适合于用有限已知样本训练建模,进而预报未知样本属性的模式识别新算法,将其用于N-(3-氧-3,4-双氢-2氢-苯并[1,4]噁嗪-6-羰基)胍类化合物的Na/H交换抑制活性类别的识别研究,用留一法考察了SVM模型的预报能力,并与Fisher判别矢量法和最近邻(KNN)法的分类预报结果进行比较,结果表明:SVM算法的预报结果优于Fisher法和KNN法的结果。因此,SVM算法可望应用于药物的构效关系研究领域。
Vapnik’s support vector machine (SVM) algorithm is applied to the pattern recognition of compound activity. The SVM algorithm is a new pattern recognition algorithm which is especially suitable for training and modeling with limited known samples and then predicting the properties of unknown samples. It is applied to N- (3-oxo-3,4-dihydro-2-hydro-benzo [ 1, 4] oxazine-6-carbonyl) guanidines were studied. The predictive power of SVM model was investigated by the one-leave method and compared with the Fisher discriminant vector method and the nearest neighbor (KNN) The results show that the SVM algorithm is better than the Fisher method and KNN method. Therefore, the SVM algorithm is expected to be applied to the field of structure-activity relationship of drugs.