基于超曲面的海量数据直接分类方法

The Large Data Direct Classifying Method Based on Hyper Surface

使用支持向量机对海量数据的分类是相当困难的 .为了解决这个问题 ,该文讨论了以下问题 :( 1)提出了一种通用的基于超曲面的直接分类方法 ,它是基于Jordan曲线定理 ,根据围绕数的奇偶进行分类判断的一种新算法 ;( 2 )提出了分类超曲面的概念 ,设计出超曲面的构造方法及基于Jordan定理的分类算法 ;( 3)对双螺旋等问题的分类实验结果说明 :分类超曲面可以有效地解决在有限区域分布很复杂的海量的非线性数据分类问题 ,并能够提高分类效率和准确率 .

It is quite difficult to classify large data by using the support vector machine. To solve the problem, several questions are discussed in this paper as below: (1) A new universal classifying method based hyper surface, HSC, is put forward based on Jordan Curve Theorem, which classify data according to whether the rewind number is odd or even. (2) The concept of separating hyper surface has been defined. Moreover, the training and classifying algorithms using HSC method are designed. (3) The experimental results of two-spiral discrimination and so on show that the separating hyper surface method can effectively solve the problem of classification of a vast amount of data and it is clear that the classifying efficiency and accuracy have been improved by using this method.