期刊名称：International Journal of Innovative Research in Computer and Communication Engineering
印刷版ISSN：2320-9798
电子版ISSN：2320-9801
出版年度：2017
卷号：5
期号：1
页码：89
DOI：10.15680/IJIRCCE.2017.0501012
出版社：S&S Publications
摘要：A many data mining problems, obtaining labels is costly and time consuming, if not practicallyinfeasible. Here implementing new unsupervised spectral ranking method for anomaly, the proposed SRA can generateanomaly ranking either with respect to the majority class or with respect to two main patterns. The spectraloptimization in Spectral Ranking method for Anomaly (SRA) can be viewed as a relaxation of an unsupervised SupportVector Machine problem. In this research work concentrate on developing the rare class kernel model, the optimizationalgorithm, and extensive computational comparisons between AUC-based and error rate based rare class nonlinearkernel learning, as well as computational efficiency improvement of RankRC over RankSVM.
关键词：Spectral Ranking method for Anomaly; Support Vector Machine RankSVM; RankRC; nonlinear kernel;learning.
