Computer Engineering and Applications ›› 2009, Vol. 45 ›› Issue (18): 144-145.DOI: 10.3778/j.issn.1002-8331.2009.18.043

• 数据库、信息处理 • Previous Articles     Next Articles

Improvement of identified adjacent object on ant colony clustering algorithm

CHEN Ying-xian   

  1. College of Business Administration,Liaoning Technical University,Fuxin,Liaoning 123000,China
  • Received:2009-02-18 Revised:2009-04-14 Online:2009-06-21 Published:2009-06-21
  • Contact: CHEN Ying-xian

蚁群聚类算法中确定相邻对象方法的改进

陈应显   

  1. 辽宁工程技术大学 工商管理学院,辽宁 阜新 123000
  • 通讯作者: 陈应显

Abstract: The basic ant colony clustering algorithm in the calculation of similarity,due to not take direction between adjacent objects,often caused by clustering algorithms do not even slow convergence.Through the introduction of the adjacent object direction angle and shielding angle,the ultimate adjacent objects are identified close to the direction of the adjacent object mask.The method to determine the adjacent object is improved.To mine the actual measurement data for the data source,using the basic ant colony clustering algorithm and the improved clustering algorithm separately.Through the experimental results of both analysis and comparison to prove that the improved algorithm enhances the clustering effect.

Key words: ant colony, clustering, adjacent object, shielding angle

摘要: 基本蚁群聚类算法在计算相似度时,由于没有考虑相邻对象之间方向的影响,往往造成聚类速度缓慢甚至算法不收敛。通过引入相邻对象方向角和屏蔽角,对方向接近的相邻对象进行屏蔽,实现对确定相邻对象方法的改进。并以矿山实际测量数据为数据源,采用基本的蚁群聚类算法和改进后的算法分别对其进行聚类。通过对这两种算法的实验结果进行分析比较,证明改进后的算法提高了聚类效果。

关键词: 蚁群算法, 聚类, 相邻对象, 屏蔽角