Method and system for data clustering for very large databases

摘要: Multi-dimensional data contained in very large databases is efficiently and accurately clustered to determine patterns therein extract useful information from such patterns. Conventional computer processors may be used which have limited memory capacity conventional operating speed, allowing massive sets processed a reasonable time with resources. The clustering process organized using feature tree structure wherein each comprises the number of points cluster, linear sum square cluster. A dense region treated collectively as single sparsely occupied regions can outliers removed tree. carried out continuously new being received processed, restructured necessary accommodate newly points.