计算机工程 ›› 2018, Vol. 44 ›› Issue (5): 19-24.doi: 10.19678/j.issn.1000-3428.0046453

• 先进计算与数据处理 • 上一篇    下一篇

基于异类近邻的支持向量机加速算法

陈景年 a,胡顺祥 b,徐力 c   

  1. a.山东财经大学 信息与计算科学系,济南 250014; b.鲁南技师学院 商务管理系,山东 临沂 276021; c.济南市公路管理局 信息科,济南250013
  • 收稿日期:2017-03-21 出版日期:2018-05-15 发布日期:2018-05-15
  • 作者简介:陈景年(1970—),男,教授、博士,主研方向为大规模数据分析、机器学习;胡顺祥,副教授;徐力,高级工程师。
  • 基金项目:
    国家自然科学基金“融合事件关系推理和情感博弈的网络不实信息演化机理研究”(61502151)。

Speeding Up Algorithm for Support Vector Machine Based on Alien Neighbor

CHEN Jingnian  a,HU Shunxiang  b,XU Li  c   

  1. a.Department of Information and Computing Science,Shandong University of Finance and Economics,Jinan 250014,China; b.Department of Commerce Management,Lunan Technician Institute,Linyi,Shandong 276021,China; c.Information Department,Jinan Road Management Bureau,Jinan 250013,China
  • Received:2017-03-21 Online:2018-05-15 Published:2018-05-15

摘要: 支持向量机的训练时间随样本增多而明显增加。为了在保持训练效果的同时提高训练速度,给出精简训练数据集的一种算法。对每个样本,通过选择异类近邻来构成训练集,利用异类近邻来选择边界样本。实验结果表明,与FCNN算法和NPPS算法相比,该算法在保持甚至增进支持向量机分类效果的同时,能大幅提高训练效率。

关键词: 支持向量机, 样本选择, k近邻, 异类, 分类

Abstract: The training time of Support Vector Machine(SVM) becomes much longer with the increase of training instances.To speeding up the training process without loosing the effect,a method for reducing training dataset is proposed.A new training set is composed with nearest neighbors of each instance selected from different classes.A boundary sample is selected by using an alien nearest neighbor.Experimental results show that compared with FCNN algorithm and NPPS algorithm,the proposed algorithm can enormously raising the training efficiency while keeping or even improving the classification accuracy of SVM.

Key words: Support Vector Machine(SVM), instance selection, k nearest neighbor, alien, classification

中图分类号: