Abstract | ||
---|---|---|
Data stream is one emerging topic of data mining, it concerns many applications involving large and temporal data sets such as telephone records data, banking data, multimedia data,…For mining of such data, one crucial strategy is analysis of packet data. In this paper, we are interested in an exploratory analysis of strategies for clustering data stream based on a sub-window approach and an efficient clustering algorithm called DCA (Difference of Convex functions Algorithm). Our approach consists of separating the data on different sub-windows and then apply a DCA clustering algorithm on each sub-window. Two clustering strategies are investigated: global clustering (on the whole data set) and independent local clustering (i.e. clustering independently on each sub-window). Our aims are study: (1) the efficiency of the independent local clustering, and (2) the adequation of local clustering and global clustering based on the same DCA clustering algorithm. Comparative experiments with clustering data stream using K-Means, a standard clustering method, on different data sets are presented. |
Year | DOI | Venue |
---|---|---|
2012 | 10.1007/978-3-642-31537-4_22 | MLDM |
Keywords | Field | DocType |
clustering strategy,data mining,different data set,global clustering,banking data,independent local clustering,dca clustering algorithm,clustering data stream,sub-window approach,efficient clustering algorithm,data stream,clustering | Fuzzy clustering,Data mining,CURE data clustering algorithm,Computer science,Artificial intelligence,Cluster analysis,Single-linkage clustering,k-medians clustering,Canopy clustering algorithm,Data stream clustering,Correlation clustering,Pattern recognition,Machine learning | Conference |
Citations | PageRank | References |
2 | 0.35 | 10 |
Authors | ||
3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Minh Thuy Ta | 1 | 6 | 1.77 |
Le An Thi | 2 | 394 | 44.90 |
Lydia Boudjeloud-Assala | 3 | 23 | 6.60 |