HSC分类法及其在海量数据分类中的应用

HSC Classification Method and Its Applications in Massive Data Classifying

使用支持向量机对非线性可分数据进行分类的基本思想是将样本集映射到一个高维线性空间使其线性可分.本文则基于Jordan曲线定理,提出了一种通用的基于分类超曲面的分类方法,简称HSC分类法,它是通过直接构造分类超曲面,根据样本点关于分类曲面的围绕数的奇偶性进行分类的一种新分类判断算法,与SVM方法相比,不需要考虑使用何种核函数,不需要做升维变换,直接解决非线性分类问题.对数据分类应用的结果说明:HSC可以有效地解决非线性数据的分类问题,并能够提高分类效率和准确度.

The main idea of SVM used for classifying nonlinear separable data is to map the data into higher dimension linear space in which the data can be separated by a hyper plane. Based on Jordan Curve Theorem,a universal classification method based on hyper surface, which is called HSC classification, is put forward in this paper. The classification hyper surface is directly made to classify massive data according to whether the wind number is odd or even. It is a novel approach that does not need to make mapping from lower dimension space to higher dimension space and considering kernel function too. It can directly solve the nonlinear multiclasses classifying problem.The test reports show that HSC can efficiently and accurately classify large data.