论文部分内容阅读
半监督多类分类问题是机器学习和模式识别领域中的一个研究热点,目前大多数多类分类算法是将问题分解成若干个二类分类问题来求解.提出两种类标号表示方法来避免多个二类分类问题的求解,一种是单位圆类标号表示方法,一种是二进制序列类标号表示方法,并利用局部学习在二类分类问题中的良好学习特性,提出基于局部学习的半监督多类分类机.实验结果证明采用了基于局部学习的半监督多类分类机错分率更小,稳定性更高.
Semi-supervised multi-class classification is a research hotspot in the field of machine learning and pattern recognition. At present, most of the multi-class classification algorithms solve the problem by decomposing the problem into several second-class classification problems.Two types of label representation are proposed to avoid multiple One is the method of label of unit circle, the other is the method of label representation of binary sequence. Based on the good learning characteristics of local learning in the second class of classification problems, a semi-supervised Classifier.Experimental results show that the semi-supervised multi-classifier based on local learning has a lower misclassification rate and higher stability.