Title :
Detecting Deviants over Data Streams
Author :
Wei, Zhang ; Zhang Wei
Author_Institution :
Sch. of Comput. Sci., Huazhong Univ. of Sci. & Technol., Wuhan, China
Abstract :
Identifying outliers is a difficult thing in data mining. We adopt the notion of deviants for outliers in data streams. Deviants are data set whose removal from the data sequence over data streams lead to sum of error SSE minimize. We present DDA algorithm to detect deviants over massive data streams. With this algorithm the histogram can more accurately determine the deviants and greatly reduce error.
Keywords :
data analysis; data mining; database management systems; least squares approximations; DDA algorithm; data mining; data sequence; data streams; deviant detection; Algorithm design and analysis; Computer science; Data mining; Educational institutions; Heuristic algorithms; Histograms; Time series analysis;
Conference_Titel :
Internet Technology and Applications, 2010 International Conference on
Conference_Location :
Wuhan
Print_ISBN :
978-1-4244-5142-5
Electronic_ISBN :
978-1-4244-5143-2
DOI :
10.1109/ITAPP.2010.5566128
