一种基于混合度的层次粒度SVM算法

发布时间：2018-09-09 14:58

【摘要】：随着现实生活中数据集规模的不断增大,设计有效的分类算法势在必行。支持向量机(Support vector machine,SVM)是一种公认的性能较好的分类算法,目前一些SVM算法是针对减少支持向量的数目来提高分类的效率。文章提出一种基于混合度的层次粒度支持向量机算法(Hierarchical Granular Support Vector Machine Algorithm based on Mixed,MHG-SVM),利用混合度对已有的层次粒度SVM分类算法进行了改进,该算法通过定义一个数据置信度和一个粒度参数挑选出重要的分类信息。从实验结果可以看出,提出的算法在处理大规模数据集方面,保持了较高的分类精度,而且支持向量机的学习和分类速度也取得了大幅度提高。
[Abstract]:With the increasing scale of data sets in real life, it is imperative to design effective classification algorithms. Support vector machine (Support vector machine,SVM) is a recognized classification algorithm with good performance. At present, some SVM algorithms are aimed at reducing the number of support vectors to improve the classification efficiency. In this paper, a hierarchical granularity support vector machine (Hierarchical Granular Support Vector Machine Algorithm based on Mixed,MHG-SVM) algorithm based on mixing degree is proposed, and the existing hierarchical granularity SVM classification algorithm is improved by mixing degree. The algorithm selects important classification information by defining a data confidence degree and a granularity parameter. It can be seen from the experimental results that the proposed algorithm maintains high classification accuracy in dealing with large data sets, and the learning and classification speed of support vector machines have been greatly improved.
【作者单位】：太原学院计算机工程系;山西大学计算机与信息技术学院;
【基金】：山西省自然科学基金(2015021096) 山西省高等学校科技创新项目(2015110)
【分类号】：TP18

【相似文献】