Dynamic Parallel Updating Algorithm For Approximate Sets Of Graded Multi-granulation Rough Set Based On Weighting Granulations And Dominance Relation

With the advancement of science and technology,data from all walks of life is exploding,and the data that computers need to process is becoming more and more complex.Faced with such a large amount of data,the traditional computing model can't handle the rapidly increasing data in an acceptable time.People's requirements for computers are also increasing day by day,requiring computers to have higher computing power for huge amounts of data.In general,traditional computers can be quite time-consuming in the process of excavating and processing data.Therefore,with the rapid increase of data volume and the rapid development of data processing technology,parallel computing emerged as a more effective data mining technology.The emergence of parallel computing is to solve the practical problems that some data volume is too large or the data volume is too complex.The parallel computing method is to transfer data between multiple parallel machines through the network,and can execute multiple instructions,multiple tasks or multiple data at the same time,which greatly shortens the running time of the data.In the face of continuous updating of big data,the traditional multi-granular rough set theory has been difficult to adapt.Therefore,when the data set is increasing,this thes' s presents a degree-based coarse-grained rough set approximation set dynamic update algorithm based on weighting granulations and dominance relation.And the related theorems of the two forms of optimism and pessimism are given.The application of the example verifies that the algorithm is effective.Parallel computing is used to parallelize serial programs to achieve parallelization of algorithms.MATLAB is used to run strings and parallel programs to compare the computational efficiency of serial and parallel programs.The experimental results show that the parallel algorithm can cope with the changes of massive dynamic update data and improve the operation efficiency when the amount of data increases.