【24h】

Classification Using Angle and Radius of Feature Vector

机译:使用特征向量的角度和半径进行分类

获取原文

摘要

In this paper, use of angle and radius information for feature space classification is proposed. The performance of the classification using either angle or the radius was evaluated on two different feature spaces for three and four-class classification problems. The results were compared with the well-known K-Nearest Neighbor (K-NN) and Naieve Bayes (NB) algorithms in terms of the ability to classify the feature space and classification time. Results show that angle and radius-based classification could generate better classification performances, especially when there are few training vectors available. Moreover, proposed methods were computationally more efficient than K-NN and NB algorithms. However, optimum combination of angle and radius-based classification is needed for developing a general classifier which will perform well in classification of different feature patterns.
机译:本文提出将角度和半径信息用于特征空间分类。针对三个和四个类别的分类问题,在两个不同的特征空间上评估了使用角度或半径进行分类的性能。在对特征空间和分类时间进行分类的能力方面,将结果与著名的K最近邻算法(K-NN)和Naieve Bayes(NB)算法进行了比较。结果表明,基于角度和半径的分类可以产生更好的分类性能,尤其是在训练向量很少的情况下。此外,所提出的方法在计算上比K-NN和NB算法更有效。然而,需要基于角度和基于半径的分类的最佳组合来开发通用分类器,该分类器将在不同特征图案的分类中表现出色。

著录项

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