Research On The Construction Of Indicator System And Comprehensive Evaluation Method Based On Rough Set

At present, there isn't systematic and normative guidance for the construction of indicator system and choice of comprehensive evaluation method; each method has its own advantages and shortages, and there is not a method that can be efficiently suitable for all kinds of indicators, objective, quantified, scientific and easy to do, thus there to bring new theory and method into the research field of comprehensive evaluation to solve the problems.Rough set theory, as a new tool of data mining, with its special attribute reduction function, attribute significance dgree, knowledge and principle obtaining, and its ability of objective and quantified data mining, can combine with comprehensive evaluation theory and to some degree can solve existing problems of comprehensive evaluation theory and methodology.This paper systematically researches on the construction and optimization of indicator system, solve the existing problems of indicator system construction and evaluation the goodness of construction; on how to apply rough set theory into the whole process of comprehensive evaluation, how to combine advantages of rough set theory with comprehensive evaluation theory and methodology, and to better solve problems of comprehensive evaluation.According to project experiences, this paper provides several common problems in the indicator system construction and provides solution way. To test the goodness of indicator system, puts forward "goodness" standard and its combination method, and optimized indicator system obtaining process based on "goodness".Based on the attribute reduction of rough set, from different aspects, this paper sets up indicator screening model based on rough set. Aiming at incomplete information of indicator system, this paper sets up null value dealing model of indicator system based on rough set. Based on attribute significance of rough set, sets up objective weight definition models and combined weight definition models based on rough set.To better apply rough set theory into comprehensive evaluation, this paper constructs comprehensive evaluation process based on rough set and methods of each step, and chooses scientific and technological credit evaluation of peer experts as an example.Rough set theory can be combined with multi other theories and methods, making use of both advantages and applying into comprehensive evaluation. Combines rough set with extenics, and sets up extensive rough set comprehensive evaluation model and uses company core competence evaluation to exemplify the method; combines rough set with fuzzy set, and sets up comprehensive evaluation method based on rough set and fuzzy set. To choose comprehensive evaluation method among many methods, this paper provides discernibility matrix model of multi comprehensive evaluation methods.Finally, to realize structuralized, method integrated and intelligent comprehensive evaluation, this paper constructs frame of comprehensive evaluation expert system based on rough set, including its function module, whole structure and data flow, and coordination of information base and reasoning engine.