• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文会议>Machine learning >K Nearest Neighbor Classification on Feature Projections
【24h】

K Nearest Neighbor Classification on Feature Projections

机译：特征投影的K最近邻分类

获取原文
获取原文并翻译 | 示例
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

This paper proposes a new approach to classification based on a majority voting on individual classifications made by the projections of the training set on each feature. We have applied the k-nearest neighbor algorithm to determine the classifications made on individual feature projections. We called the resulting algorithm k-NNFP, for k-Nearest Neighbor on Feature Projections. The most important characteristic of this technique is that the training instances are stored as their projections on each feature dimension. This allows the classification of a new instance to be made much faster than k-NN algorithm. The voting mechanism reduces the negative effect of possible irrelevant features in classification. Also the classification accuracy of k-NNFP increases when the value of k is increased, which indicates that the process of classification can incorporate the learned classification knowledge better when k increases. The k-NNFP algorithm is compared with the k-NN algorithm, in terms of classification accuracy and running time on some real-world and artificial datasets.
机译：本文提出了一种新的分类方法，该方法基于对每个特征的训练集的投影对各个分类进行的多数表决。我们已应用k最近邻算法来确定对单个特征投影进行的分类。我们将生成的算法k-NNFP用于特征投影的k最近邻。该技术最重要的特征是训练实例作为它们在每个特征维上的投影存储。这使得新实例的分类比k-NN算法快得多。投票机制减少了分类中可能不相关的特征的负面影响。当k的值增加时，k-NNFP的分类精度也增加，这表明当k增大时，分类过程可以更好地结合学习到的分类知识。在分类精度和在某些实际数据集和人工数据集上的运行时间方面，将k-NNFP算法与k-NN算法进行了比较。

著录项

  • 来源
    《Machine learning》|1996年|12-19|共8页
  • 会议地点 Bari(IT);Bari(IT)
  • 作者

    Aynur Akkus; H. Altay Guevenir;

  • 作者单位

    Dept. of Computer Engr. and Info. Sci. Bilkent University 06533 Ankara, Turkey;

    Dept. of Computer Engr. and Info. Sci. Bilkent University 06533 Ankara, Turkey;

  • 会议组织
  • 原文格式 PDF
  • 正文语种 eng
  • 中图分类 计算机的应用;
  • 关键词

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号