The Study Of Several Problems In The Application Of Rough Set Theory

Rough set theory, as a forceful tool in processing uncertainties, has been applied in wide areas. The dissertation focuses on data pretreatment and reduction of decision tables and proposes some meaningful methods and improvements. The research works in the thesis can be described as follows:1.The dissertation redefine the key concepts of basic rough set model by relation algebra to build the identity between rough set and relation database, so that lays a solid foundation for the high effective processing of decision tables to take advantage of set computations.2.To the processing of inconsistent decision tables, time factor of process is introduceded. Two strategies , Trend analysis is introduced to find it, then verify applicability of the first method..3. After summarize present discretization algorithms of continuous variables, then improve the discretization algorithm of continuous variables based on Boolean logic. The combinations of algorithms are attempted and experiments prove the technique is an effective way to achieve better results.4.The general reduct algorithm was improved and a new heuristic one was introduced.5.T0 the disposal of super large data set, the building of cluster systems is put forward and the algorithms of task disassemble and computation are designed.6. An example of the application of rough set theory is given. Rough set theory is applied in queue system simulation to compress information and extract decision rules from decision table that comes from original records. These rules reflect the factors that affect the arrival rate of customers. The new simulation process is carried out based on above results. The new way enhances the value of the output of simulation to decision making.Finally, a recapitulative conclusion is given, and the future research directions are proposed.