交遇区样本分类的应用

在模式识别中，许多问题是非线性的．对于未知的样本需要按属性来进行分类，并且由于空间条件的复杂性高，分类器的设计方法也有很多种．利用“交遇区”中的样本的特殊性，把非线性的问题转换成分段线性问题来处理，并设计了基于“交遇区”的样本分段线性分类器，来对未知的样本进行分类．该分类器可以应用于数据挖掘、模式识别、人工智能等领域．

The application of the classification of samples in overlapping region

In pattern recognition, many problems are nonlinear. Many unknown samples need to be classified by their attributes. Because of the high complexity of spatial conditions, there are many ways for designing classification. In this paper, using the particularity of samples in overlapping region, nonlinear problems are changed into many segmentalized linear problems. And a linear segmentation classification for samples in overlapping region is designed. Unknown samples can be classified by the classification. The classification can be applied in data mining, pattern recognition and artificial intelligence.