一种针对不均衡数据集的SVM决策树算法  

An SVM Decision Tree Algorithm for Unbalanced Data Sets

作　　者：黄勇 魏乐 HUANG Yong;WEI Le(College of Computer Science,Chengdu University of Information Technology,Chengdu 610225,China;College of Software Engineering,Chengdu University of Information Technology,Chengdu 610225,China;Sichuan Key Laboratory of Software Automatic Generationand Intelligent Services,Chengdu 610225,China)

机构地区：[1]成都信息工程大学计算机学院,四川成都610225 [2]成都信息工程大学软件工程学院,四川成都610225 [3]软件自动生成与智能服务四川省重点实验室,四川成都610225

出　　处：《成都信息工程大学学报》2019年第3期274-277,共4页Journal of Chengdu University of Information Technology

基　　金：四川省科技计划重点研发项目资助(2017GZ0309);四川省教育厅青年基金重点资助项目(16ZA0208)

摘　　要：针对文本分类问题中常遇到的数据分布不均的情况,提出一种新的SVM决策树算法。算法在构造分类器结点时,运用动态规划的思想,寻找类别数和样本数量同时最优的分配方案。实验结果表明,该方法比基于完全二叉树的SVM分类器准确率有明显提升。A new SVM decision tree algorithm is proposed for the uneven distribution of data commonly encountered in text classification problems. When constructing the nodes of classifier, the algorithm uses the idea of dynamic programming to find the optimal allocation scheme of both the number of categories and the number of samples.The experimental results show that the proposed method has a significantly better accuracy than the SVM classifier based on the complete binary tree.

关 键 词：大数据 自然语言处理 动态规划 完全二叉树 支持向量机 文本分类 机器学习 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]