Title :
CBLOS: Improving local outlier detection
Author :
Mi, Hongjuan ; Wang, Jikui
Author_Institution :
Information Engineering School, Lanzhou Commercial College, Lanzhou 730020, PR China
Abstract :
In this paper, we propose a new technique for local outliers detection based on clustering. In terms of the local data behavior, we construct a formula for computing cluster-based local outlier score (i.e. CBLOS) within clusters. Using real-world datasets, we demonstrate that CBLOS can be used to efficiently identify outliers.
Keywords :
Algorithm design and analysis; Breast cancer; Clustering algorithms; Complexity theory; Data mining; Educational institutions; Recurrent neural networks; clustering; data mining; outlier detection;
Conference_Titel :
E -Business and E -Government (ICEE), 2011 International Conference on
Conference_Location :
Shanghai, China
Print_ISBN :
978-1-4244-8691-5
DOI :
10.1109/ICEBEG.2011.5881753
