摘要
近年来互联网数据规模呈爆炸式增长,如何对大数据进行分析已成为热门话题。然而,采集的数据很难直接用于分析,需要进行一定程度的预处理,以提高大数据质量。通过使用分裂式的迭代过程,可以逐步将数据集分裂为子集,避免了传统聚类算法聚类开始时需要确定集群数的限制,并降低了算法的时间复杂度。此外,通过基于阈值的噪声数据过滤,可以在迭代过程中剔除噪音数据,提升了聚类算法对脏数据的忍耐力。
In recent years,the scale of internet data has explosive growth,which makes big data analysis become a hot topic.However,it is difficult to directly utilize the collected data,so a certain degree of pretreatment had to be made in order to improve the quality of big data.In this work,the data set will be gradually divided into smaller subsets by using the split iterative process,which can effectively avoid the limitation of traditional clustering algorithm and reduce the time complexity.In addition,by thresholdbased noise data filtering,the dirty data can be eliminated during the iterative process so as to enhance the tolerance of the clustering algorithm to the dirty data.
出处
《华东理工大学学报(自然科学版)》
CSCD
北大核心
2017年第6期849-854,862,共7页
Journal of East China University of Science and Technology