Title :
n-INCLOF: A dynamic local outlier detection algorithm for data streams
Author :
Gao, Ke ; Shao, Feng-Jing ; Sun, Ren-Cheng
Author_Institution :
Coll. of Inf. Eng., Qingdao Univ., Qingdao, China
Abstract :
With the development of the data stream technology, The way to detect anomalies in data streams accurately has been widespreadly concerned. According to the problem that the distribution of the number of the outliers in data streams is unstable, in this paper, the n-IncLOF incremental outlier detection algorithm is proposed which could adjust the n-threshold automaticly. The experiment of oultlier detection of the data stream proves that n-IncLOF algorithm could adjust to the change of the number of outliers effectively and it not only improves the detection rate greatly but also lowers the false alarm rate compared to the original incremental algorithm.
Keywords :
data mining; pattern classification; anomalies detection; data stream technology; dynamic local outlier detection algorithm; incremental algorithm; n-IncLOF; n-threshold; Algorithm design and analysis; Complexity theory; Data mining; Data models; Detection algorithms; Heuristic algorithms; Signal processing algorithms; Data Mining; Data Streams; Outlier; n-threshold;
Conference_Titel :
Signal Processing Systems (ICSPS), 2010 2nd International Conference on
Conference_Location :
Dalian
Print_ISBN :
978-1-4244-6892-8
Electronic_ISBN :
978-1-4244-6893-5
DOI :
10.1109/ICSPS.2010.5555276
