원문정보
초록
영어
To solve the clustering algorithm based on grid density on uncertain data stream in adjustment cycle for clustering omissions, the paper proposed an algorithm, named GCUDS, to cluster uncertain data steam using grid structure. The concept of the data trend degree was defined to describe the grade of a data point belonging to some grid unit and the defect of information loss around grid units was removed in the GCUDS algorithm. The GCUDS algorithm obtained better results of clustering and higher time efficiency than other algorithms over uncertain data stream, through improving the traditional online clustering framework and maintaining three buffers of micro-cluster. Experimental results showed that the GCUDS algorithm could effectively cluster in different shape database and outperform existing methods in clustering quality and efficiency.
목차
1. Introduction
2. The Related Model and Definitions
2.1. Uncertain Data Stream Model
2.2. Network Model
2.3. Related Definition of Grid Processing Mechanism
3. Clustering Process of Uncertain Data Stream
3.2. Evolutionary Function of the Cluster-CkeckClustersProcess()
3.3. The Processing Function of New Coming Data--UpdateGrid()
3.4. Cluster Selection Function-FindOptimalCluster()
3.5. Time Complexity
4. Experimental Results and Analysis
4.1. Test Data Design and Parameter Setting
4.2. Analysis of Clustering Effect
4.3. Analysis of Clustering Time
4.4. The Scalability of GCUDS
5. Conclusion
Acknowledgements
References