Efficient updating reduction based on positive region for distributed data

With a rapid growth of network and sensor technology, the sources of mining distributed data becomes one of vital task in knowledge discovery. Rough Set theory(RST) is a valid tool that can dispose of inconsistent and incompletion data. Computing reduct is an important step for the rule extraction in RST. Furthermore, the data changes with time in distributed systems, and updating the reduct of distributed data is a key problem in knowledge discovery. In this pursuit, the present study envisaged a matrix-based method for computing reduct of the data distribution based on positive region. Subsequently, an incremental reduction approach was proposed to update the reduct of distributed data when some objects dynamically changed in the distributed system. The results in-dicated that the incremental reduction approach exhibited preferable feasibility. Multiple experiments testified the effectiveness of the given incremental approach for distributed data.