Title :
An Efficient Algorithm for Finding Frequent Items in a Stream
Author :
Tu, Li ; Chen, Ling ; Zhang, Shan
Author_Institution :
Dept. of Comput. Sci., Nanjing Univ. of Aeronaut. & Astronaut., Jiangyin, China
Abstract :
Most of the existing algorithms for mining frequent items over data streams do not emphasis the importance of the more recent data items. We present an efficient algorithm where a fading factor lambda is used for computing frequency counts exceeding a user-specified threshold over data streams. Our algorithm lambda-Miner can detect epsiv-approximate frequent items of a data stream using O(epsiv-1) memory space and the processing time for each data item is O(1). Experimental results on several artificial data sets and real data sets show that lambda-Miner performs better than lambda-LC in terms with precision, memory requirement and time cost.
Keywords :
data analysis; data mining; data analysis; data stream; frequent item mining; lambda fading factor; user-specified threshold; Computer science; Costs; Counting circuits; Data mining; Electronic commerce; Fading; Frequency estimation; Intrusion detection; Sampling methods; Software algorithms; data mining; data stream; fading factor; frequent items;
Conference_Titel :
Electronic Commerce and Security, 2009. ISECS '09. Second International Symposium on
Conference_Location :
Nanchang
Print_ISBN :
978-0-7695-3643-9
DOI :
10.1109/ISECS.2009.188
