Download now Free registration required
Effective data analysis using multiple databases requires highly accurate patterns. Local pattern analysis might extract low quality patterns from multiple large databases. Thus, it is necessary to improve mining multiple databases using local pattern analysis. The authors present existing specialized as well as generalized techniques for mining multiple large databases. They formalize the idea of multi-database mining using local pattern analysis and propose a new generalized technique for mining multiple large databases. It improves the quality of synthesized global patterns significantly. They conduct experiments on both real and synthetic databases to judge the effectiveness of the proposed technique.
- Format: PDF
- Size: 259.3 KB