Incomplete Information System's Knowledge-Gaining Method Studies

Today,with the rapid development of computer and network information technology,the information and data from the various fields are increased dramatically.Because of the involvement of human beings,the data and information become uncertainty in the more significant,and the relationship of information and data becomes more and more complex.How from massive,chaotic,the strong jamming incomplete information to unearth latent,novel,correct,the valuable knowledge,this will give the intelligence information processing to propose the stem challenge,has had an artificial intelligence area research brand-new domain-data mining from this(DM) and the database knowledge discovery(KDD).In DM and in KDD many methods,the formal concept analysis,the fuzzy clustering analysis,the rough set and so on regarding the processing incomplete information are the effective methods.Fuzzy clustering analysis and rough set are important method of the expression and the processing uncertainty data,they not only may process the incomplete data,the noise or the inaccurate precise information,moreover in the development data's uncertainty model aspect is useful,can provide more flexible and smooth performance than the traditional methods,can provide the conventional routes is more nimble than,a smoother performance.Formal concept analysis and rough set are used for data modeling and data analysis as mathematical tools,formal concept analysis has strong algebraic structure,it has an advantage to obtain clearer and more logical knowledge than other data analytical tools,and rough set has been made a lot of research results for all kinds of complicated data sets of knowledge acquisition.Since the research target of formal concept analysis and rough set has a natural similarity,the fusion research between them will expand the handling capacity of rough set in the incomplete information system and become one of the keys to push the rough set more practical in the further.This article tracks the international academic front,based on the incomplete formal context,has established a concept lattice structure model and knowledge library,has presented one kind newly based on the fuzzy clustering analysis information complete method;has discussed the fusion research between the formal concept analysis and rough sets initially.The conclusion of this paper not only provides a new way for the information completing,but also has the significant application value.This article contains the following major achievements:(1) Knowledge gains in the incomplete formal context by formal concept analysis.The paper defines object-attribute consistent granules in the incomplete formal context,combines the theory of formal concept analysis to present a new method,namely,the information acquisition method of object-attribute consistent granule-based.In the process of information acquisition,on the one hand,we refer to the expert advices in the relative realm and build knowledge base,so it ensures the reliability of information acquisition,on the other hand,the idea of granular computing is introduced, which will make the problems be solved abstractly or simply under the different granules,so the problems will be understood easily,but not make us submerge in the unnecessary details of the problems,evidently,it cut down the complication of information acquisition.Examples prove the method which is effective,based on the incomplete formal context,the paper not only establishes a concept lattice structure model and knowledge library,but also presents a new theory frame for the analysis and managing data.(2) Using the fuzzy clustering theory in the incomplete information system.This paper constructs a new managing data platform by using the fuzzy clustering theory to deal with the incomplete information system, namely,consistent fuzzy granular frame system.The system provides data support under the different granularity to restore the missing data correctly,it contain two parts,which are consistent fuzzy granular frame families and the set of missing information granules,and we will use the consistent fuzzy granular frame families to restore the missing date in the missing information granules.In the process of missing data acquisition under multi-granularity, we consider the confine of the granularity and the effecting of the different granularity should satisfy the law of diminishing marginal utility,because of the descending of the effecting of the granularity when it is larger.Fuzzy synthetic evaluation was introduced to measure the amount of the missing datum,since the amount of the missing datum may affect the parameters setting.Example proves the method is effective,Compared with some conventional methods,it can respond humanity's cognition process.(3) The fusion research between the formal notion analysis and the rough set.As mathematical instruments,the formal notion analysis and rough sets are used for the data modeling and analysis,the formal notion analysis have the strong algebraic structure,it has an advantage to obtain clearer and more logical knowledge than other data analytical tools,and rough set has been made a lot of research results for all kinds of complicated data sets of knowledge acquisition.This article has realized the formal concept analysis and the rough set fusion research,portrays the dependent space,the attribute reduction,the independence,the nucleus and the function dependence and so on rough centralism core concept with the formal notion analysis.This theory is simple and advantageous for the understanding,is helpful carries on a deeper level of the exploration and research.