基于覆盖算法的分类器的设计与应用

覆盖算法和支持向量机是两种重要的机器学习分类方法,但长期以来一直缺少基于覆盖算法的通用分类器,在一定程度上阻碍了覆盖算法的推广.论文设计和实现了基于覆盖算法的通用分类器J Cover,该分类器有友好的人机交互界面,能对数值型数据集进行有效分类,并给出详细的实验结果.通过与LIBSVM的对比实验表明J Cover在识别率、泛化能力和性能上具有优越性.

Design and application of the classifier based on covering algorithm

Covering algorithm and support vector machine are two important classifications of machine learning.But there was not a general-purpose classifier based on covering algorithm for a long period of time,so it hindered the promotion of covering algorithm to some extent.A classifier,J cover,based on the covering algorithm was designed and applied in the paper.It had a friendly interactive interface,and could make a good classification and give the detail of the experimental results.By comparing with LIBSVM,the results of the experiments proved that J Cover had a good classifying accuracy,extensibility and performance.