METHOD AND APPARATUS FOR FINDING CLUSTER IN DATA STREAM AS INFINITE DATA SET HAVING DATA OBJECTS TO BE CONTINUOUSLY GENERATED

摘要

Disclosed is a method and apparatus for finding a cluster in a data stream as an infinite data set having data elements, which are continuously generated. A method of finding a cluster in a data stream according to an embodiment of the present invention includes the steps of: (a) updating statistical distribution information of a grid-cell corresponding to a currently generated data element among the grid-cells, statistical distribution information on previously generated data elements being managed using grid-cells, which are partitioned within the range of a data space and have statistical distribution information of data elements within the range; (b) comparing the occurrence frequency of the data element in the grid-cell according to the update result of the statistical distribution information with a predefined partitioning threshold, partitioning the grid-cell into a plurality of grid-cells according to the comparison result, and estimating statistical distribution information of the partitioned grid-cells; (c) recursively performing the step (a) or (b) until the grid-cell becomes a unit grid-cell having a predefined size; and (d) comparing the occurrence frequency of a data element in the unit grid-cell with a predefined minimum support and defining a set of a plurality of unit grid-cells as a cluster according to the comparison result.