基于衰减滑动窗口数据流聚类算法研究

数据流具有数据流量大、流量连续且快速、难以存储和恢复等特性,其挖掘质量和效率是检验挖掘算法的重要标准.传统的数据流聚类挖掘算法是基于界标窗口、滑动窗口和衰减窗口模型,其算法的聚类质量较差,时间复杂度高等不足,就此类问题,研究一种滑动衰减窗口的数据流聚类算法,并对算法进行了设计与实现,有效的改善传统数据流算法聚类质量和时间效率的问题.仿真实验结果表明了该算法的有效性,达到了较满意的效果.

Attenuation data streams based on sliding window clustering algorithm

Data streams with continuous and rapid features,it is difficult to storage and recovery,and its mining quality and efficiency are the important standard of test mining algorithm.The traditional data stream clustering algorithm is based on the landmark window,sliding window and the attenuation window model,the clustering algorithm of poor quality,high time complexity,on a sliding attenuation window data stream clustering algorithm,and the algorithm design and implementation,effectively improve the traditional data flow algorithm the clustering quality and efficiency problem.Through the simulation experiment,verified the effectiveness of the algorithm,to achieve a more satisfactory effect.