一种新的海量数据分类方法

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。文章则基于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,不需作升维变换,不需要考虑使用何种核函数,而直接地解决非线性分类问题。对数据分类应用的结果说明:基于分类超曲面的分类法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。

A Novel Classification Method in Large Data

The main idea of classifying nonlinear separable data by using Support Vector Machine is to map the data into higher dimension linear space in which the data can be separated by hyper plane.Based on Jordan Curve Theo-rem,a universal classification method based on hyper surface is put forward in this paper.The classification hyper sur-face is directly made to classify huge data according to whether the wind number is odd or even.It is a novel ap-proach that need not make mapping from lower dimension space to higher dimension space and need not consider ker-nel function too.It can directly solve the nonlinear classify problem.The experimental reports show that the new method can efficiently and accurately classify huge data.