Knowledge Reduction And Imputation Based On Granular Computing In Formal Concept Analysis

Formal concept analysis(FCA)is a kind of data processing tool which can effectively formalize management and analysis of data.As its main data structure,concept lattice shows the relationship between formal concepts.Since the granular computing was put forward in 1997,the method of using particle size to solve the problem is presented in different degrees in all major disciplines.In reality,for example human factors,equipment failure,suspected privacy and other reasons often result in the non-ideal storage of database information.the consequent incomplete redundant data,not only increase the difficulty that the database information management personnel and knowledge acquisition accuracy,and the error operation of data or data processing method is too inefficient,it is entirely possible to individuals and the whole enterprise caused irreparable damage.For better and faster use of these data,data is often preprocessed to improve the overall availability of the database.This thesis mainly aims at the integration of formal concept analysis and granular computing for research,through the grain of the thoughts of finished filling the incomplete decision formal context,at the same time to formal context of knowledge reduction was studied,the specific research work are as follows:1.To solve the problems such as the deficiency of the algorithm in the thought and the low efficiency of filling the algorithm.Firstly,the upper and lower bounds of approximation degree between objects are introduced,and the fuzzy approximate evaluation matrix is constructed.Then,combined with the quotient space granular computing model,through stratified hierarchical structure model,the granulation of the object set,under the quotient set for the original problem,put forward a kind of missing data filling algorithm based on quotient space,and through the real example is given full fill the results.In the real data set,we verified that the algorithm can fill the data with high data loss rate and maintain high accuracy.2.A reduction method based on Rough logic formula is proposed to solve the problem of large calculation,low efficiency and unreasonable results of the reduction algorithm in the formal context.This method uses the grain derived from Rough logic formula to obtain a subset of object ? degree satisfied by the grain,and then the coverage reduction of granular sentence is carried out on this basis.Finally,through the comparison between the actual example and other reduction algorithms,the high efficiency of the proposed algorithm is verified,and the reduction result is moreprecise and reasonable.