Study On Rough Set Theory And Its Application In Relational Database

By rapid development of the computer technology and explosive rise in data, the classical methods of knowledge discovering can not meet people's demand for the knowledge hidden in the data. People are searching for a data analytic method that is speedy, available and intellective. Rough sets theory is a new mathematical tool, which analyses the facts hiding in data without any additional knowledge about the data, and a pithily tool for processing vague, noisy and uncertain knowledge. In this paper, the basic concepts of rough sets theory and research actuality of combining rough sets theory and relational database is introduced, the efficient algorithm of attributes reduction, correlative definition and algorithm of reduction in database and the rough relational database model is presented.It is important to seek effective reduction algorithm in rough sets because of less efficiency of the complete algorithm of attributes reduction based on the discernibility function and incomplete of heuristic reduction algorithm. In this paper, considering the low efficiency of the reduction algorithm, a complete heuristic reduction algorithm is presented The better effect can be tested by the experiment on UCI data.Generally, the research of knowledge discovery is based on relational database, database reduction algorithm based on relational database directly obtains core of database, attributes reduction and rule reduction by operator in relational database. It is a simple and effective.Essentially, rough relational database model in this paper, which extending classical rough sets model, is a multi-valued information system. The definition of Multi-valued information system is extended and the spatial structure of rough relational database model is analyzed in this paper. The measurement of uncertainty in rough database is provided. By analysis of the spatial structure of RRDB, the upper approximation, the lower approximation and value-reduced is educed.It provides the foundation for more studies in rough relational database model.