本文提出了一种将支持向量机分类和最近邻分类相结合的方法 ,形成了一种新的分类器 .首先对支持向量机进行分析可以看出它作为分类器实际相当于每类只选一个代表点的最近邻分类器 ,同时在对支持向量机分类时出错样本点的分布进行研究的基础上 ,在分类阶段计算待识别样本和最优分类超平面的距离 ,如果距离差大于给定阈值直接应用支持向量机分类 ,否则代入以每类的所有的支持向量作为代表点的K近邻分类 .数值实验证明了使用支持向量机结合最近邻分类的分类器分类比单独使用支持向量机分类具有更高的分类准确率 ,同时可以较好地解决应用支持向量机分类时核函数参数的选择问题 In this paper, a new method of combining SVM classification and nearest neighbor classification is proposed, which forms a new classifier.Firstly, the analysis of SVM shows that it is equivalent to only one At the same time, based on the research of the distribution of error sample points in SVM classification, the distance between the sample to be identified and the super-plane of the optimal classification is calculated in the classification stage. If the distance difference is greater than the given threshold Direct application of support vector machine classification, or substitution into all support vectors for each class as a representative of the K-nearest neighbor classification.Numerical experiments show that the use of support vector machines combined with nearest neighbor classification classifier classification than the use of support vector machine classification alone has more High classification accuracy, at the same time can be a better solution to the problem of selection of kernel function parameters when using SVM classification