基于kmeans-SVM的二叉树粗分类方法

针对大型数据库中进行匹配识别时存在识别速度慢、时间长、影响实时应用效果的问题,提出了一种树形层次结构的粗分类方法。通过k-means得到两类粗分类的样本,用这两类粗分类数据训练SVM分类器,找到分类超平面,再不断调整分类超平面,最后构建二叉树型结构达到粗分类的目的。三个方法相结合很好地缩小目标的搜索范围,提高了识别时候的效率。

Rough Classification Method Based on Binary Tree and Kmeans-SVM

Aiming at the problem of matching recognition in large database,the recognition speed is slow,the time is long,and the effect of real-time application is affected.A rough classification method of tree hierarchy is proposed.Two types of rough classification samples are obtained by k-means,then the SVM classifiers are trained by these two types of rough classification data to find the classification hyperplane,and the classification hyperplane is continuously adjusted.Finally,the binary tree structure is constructed to achieve the coarse classification.Combining these three methods for coarse classification processing achieves good re?sults and improves the efficiency at the time of recognition.