摘要
孤立点挖掘在电子商务犯罪,信用卡欺诈等领域已有成功应用。本文尝试使用CLARANS聚类算法进行孤立点探测,并将其应用于大型数据集数据的正确性检验。实验表明该方法能够克服常规数据抽查方法的不足。
Outlier detection has been very successful in applications such as EB-criminal, credit cheat etc. In this paper we use CLARANS for outlier detection and apply it to the accuracy test of data in large datasets. Experiments show that this method can avoid deficiencies of conventional spot check.
出处
《北京电子科技学院学报》
2007年第4期81-83,80,共4页
Journal of Beijing Electronic Science And Technology Institute
基金
甘肃省自然科学基金项目(3ZS051-A25-045)
