计算机应用 ›› 2019, Vol. 39 ›› Issue (9): 2494-2498.DOI: 10.11772/j.issn.1001-9081.2019020299

• 人工智能 • 上一篇    下一篇

面向概念漂移问题的渐进多核学习方法

白东颖1,2, 易亚星1, 王庆超1, 余志勇1   

  1. 1. 火箭军工程大学 作战保障学院, 西安 710025;
    2. 空军工程大学 防空反导学院, 西安 710051
  • 收稿日期:2019-02-27 修回日期:2019-04-29 出版日期:2019-09-10 发布日期:2019-05-14
  • 通讯作者: 易亚星
  • 作者简介:白东颖(1982-),女,陕西宝鸡人,讲师,硕士,主要研究方向:智能信息处理;易亚星(1966-),男,湖南岳阳人,教授,博士,CCF会员,主要研究方向:系统建模与仿真;王庆超(1988-),男,河北清河人,讲师,博士,主要研究方向:遥感图像分析;余志勇(1972-),男,湖北恩施人,教授,博士,主要研究方向:电磁兼容。
  • 基金资助:

    国家自然科学基金青年项目(61806219)。

Gradual multi-kernel learning method for concept drift

BAI Dongying<sup>1,2</sup>, YI Yaxing<sup>1</sup>, WANG Qingchao<sup>1</sup>, YU Zhiyong<sup>1</sup>   

  1. 1. Institute of Combat and Support. Xi'an Institute of High-Tech, Xi'an Shaanxi 710025, China;
    2. Institute of Air and Missile Defense, Air Force Engineering University, Xi'an Shaanxi 710051, China
  • Received:2019-02-27 Revised:2019-04-29 Online:2019-09-10 Published:2019-05-14
  • Supported by:

    This work is partially supported by the National Natural Science Foundation of China for Young Scholars (61806219).

摘要:

针对概念漂移问题,构建数据特性随时间发生渐进变化特点的分类学习模型,提出一种基于渐进支持向量机(G-SVM)的渐进多核学习方法(G-MKL)。该方法采用支持向量机(SVM)为基本分类器,进行多区间上的子分类器耦合训练,并通过约束子分类器增量方式使模型适应数据渐进变化特性,最终将多个核函数以线性组合方式融入SVM求解框架。该方法综合发挥了各个核函数的优势,大大提高了模型适应性和有效性。在具有渐变特性的模拟数据集和真实数据集上将所提算法与多种经典算法进行了对比,验证了所提算法在处理非静态数据问题的有效性。

关键词: 概念漂移, 支持向量机, 多核学习, 子分类器, KKT条件

Abstract:

Aiming at the concept drift problem, a classification learning model with the characteristics of data changing progressively over time was constructed, and a Gradual Multiple Kenerl Learning method (G-MKL) based on Gradual Support Vector Machine (G-SVM) was proposed. In this method, with Support Vector Machine (SVM) used as the basic classifier, multi-interval sub-classifier coupling training was carried out and the incremental method of constraining sub-classifier was used to adapt the model to the gradual change of data. Finally, multiple kernels were integrated into SVM solution framework in a linear combination manner. This method integrated the advantages of different kernel functions and greatly improved the adaptability and validity of the model. Finally, the comparison experiments between the proposed algorithm and several classical algorithms were carried out on the simulated and real datasets with gradual characteristics, verifying the effectiveness of the proposed algorithm in dealing with non-stationary data problems.

Key words: concept drift, Support Vector Machine (SVM), multi-kernel learning, sub-classifier, Karush-Kuhn-Tucher (KKT) condition

中图分类号: