| With the rapid development of information technology and the continuous progress of information processing technology,information and data are redundant,scalable,and complex.In the face of these irregular,massive and rapidly updated data,how to mine the knowledge hidden in these data and conduct effective data analysis is one of the most noteworthy research topics in the era of big data.It is also an opportunity and challenge for researchers in the field of artificial intelligence.As a powerful tool for data mining and knowledge discovery,rough set theory proposed by Pawlak focuses on the approximation of lower approximation and upper approximation,and characterizes uncertain information with the known information.In order to describe the fuzziness and uncertainty of data in information systems more accurately,K.T.Atanassov proposed the concept of intuitionistic fuzzy sets.Different from traditional fuzzy sets,intuitionistic fuzzy sets both consider membership degree,non-membership degree and hesitation degree when dealing with fuzzy information,which contains more complete and comprehensive information than traditional fuzzy sets.Due to the complementarity of intuitionistic fuzzy sets and rough set theory in dealing with uncertainty and fuzzy problems,the combination of intuitionistic fuzzy set and rough set theory has attracted the attention of scholars at home and abroad.Its related research results have been applied in data mining,decision analysis,medical diagnosis and other important fields.Under the background of intuitionistic fuzzy information system,this paper studies the neighborhood dominance rough set model and the adjustable-perspective rough set model based on the classical rough set model,and combine the multi-granulation theory.The research also includes the martix-based dynamic approximation update method and the rule extraction approach.Finally,the roughness algorithm,dependency algorithm and matrix-based dynamic updating algorithm of various models are designed,and the effectiveness of the proposed method is verified by a large number of experiments.The main innovations are as follows:1.This thesis integrates the neighborhood relation and the dominance relation in intuitionistic fuzzy ordered information system,the neighborhood dominance rough set is constructed.Meanwhile,this research establishes the multigranulation neighborhood dominance rough set(MNDRS),and discuss related connections and properties between NDRS and MNDRS.A series of experiments are implemented to illustrate the feasibility and effectiveness of the proposed models.2.We study the multi-granulation neighborhood dominance rough set matrix-based updating method when the universe is dynamic in intuitionistic fuzzy ordered information system.Furthermore,the corresponding matrix dynamic algorithm is designed for data increase or decrease.Finally,this thesis tests 8 data sets from UCI,and designs comparative experiments to evaluate the matrix dynamic algorithm.With a view to the computational efficiency,the experimental results show that matrix-based method has obvious advantages.3.In static intuitionistic fuzzy ordered information systems,this study uses triangular norm,uncertainty measures and rough set model as tools to investigate the new adjustableperspective dominance relation,and then simplifies the complex ranking problem.Singleperspective rough set model(SPRS)and different-perspective rough set model(DPRS)are proposed based on the adjustable-perspective dominance relation.Additionally,SPRS and DPRS are compared with the previous dominance rough set models by the indexes of order classification accuracy,roughness and dependence.Finally,the feasibility and superiority of the proposed model are verified by experiments on UCI data sets. |
PDF Full Text Request