Research On Incremental Factor Reduction Algorithm Based On Factor Space Theory

In order to improve the intelligence of factor databases,accelerate the speed of factor reduction,solve the problem that the static factor reduction can not adapt to the dynamic change of database and the dynamic change of factors in the factor analysis table,the reduction algorithms of adding new objects and new factors in the data set are proposed,respectively.When new objects are added,this paper studies the update recurrence formula of the decision degree and designs dynamic factor reduction algorithm for increasing objects by analyzing the change of decision domain.When new factors are added,based on effectively using the results of static factor reduction on the original factor analysis table,the dynamic factor reduction algorithm is presented by analyzing the relation between the decision degree of the increasing factors and the decision degree of factors in the original reduction factor sets.This paper selects some data sets from UCI database to do numerical experiments based on the two algorithms respectively,and the results show that the dynamic factor reduction algorithm is effective and feasible.Considering the factor diverging problem in the factor analysis and factor reduction process,a factor selection method is established,which can reduce the number of inference rules and make the rules more adaptable.