Research Of Uncertain Information System And Decision Based On Rough Set Theory

With the application of the emerging information technology such as cloud computing and big data, the data from all fields grows rapidly in which the structured data is still one of the main form of data expression. Much redundant and uncertain data often exist which reduce the processing capacity of pattern classification as well as the discernibility power of decision-making. As two forms of expression for uncertain and incomplete information, the interval-valued data and intuitionistic fuzzy data are two important structured data. How to find valuable information and rules from these two types of uncertain data so as to provide decision-making reference for the manager is still be one of research hot spots in the field of management and decision science.As one of the most important methods in the field of data mining,the most significant advantage of rough set is that it can achieve classification and decision rule acquisition of the data without providing priori knowledge for solving the problems except the information provided by the data itself. The rough set theory has been successfully used in machine learning, data mining, decision analysis and many other fields. Because the classical rough set is based on the equivalence relation which is too strict, there are many limitations in dealing with the uncertain data. So there is no doubt that the extension for the classical rough set model has important significance for knowledge reduction and decision rules acquisition of uncertain information systems.Based on rough set and combined with the present domestic and abroad research,the attributes reduction as well as decision rules acquisition problems in interval-valued information system and intuitionistic fuzzy information system under single granulation and multi-granulation background have been researched systematically in this paper. At the same time,a novel group decision-making attribute rough set model is constructed and applied for the traffic accident factor correlation analysis.The main innovations of this paper are as follows:(1) Based on the analysis of the existing tolerance relation deficiency for the clustering of interval-valued information, the rough set model based on fuzzy equivalence relation in interval-valued information system from point of single granulation and multi-granulation is proposed in this paper. The discernibility matrix as well as judging theorem and method of attributes reduction in interval-valued information systems are proposed. Whats more, the certainty factor as well as the support theorem and the acquisition method of decision rules in the interval-valued information system based on fuzzy equivalent relation are then given.(2) The rough set model in intuitionistic fuzzy information system are proposed from point of single granulation and multi-granulation . The preference relation as well as discernibility matrix are defined and the effective attributes reduction method is also proposed . The relative reduction calculation method based on the classification quality of intuitionistic fuzzy decision system is then given. At the same time, the relative attribute importance as well as the order extraction method are then given. Whats more,the optimistic model as well as pessimistic model of intuitionistic fuzzy information system are proposed and then not only their properties are analyzed but also the differences and relationships between single-granulation and multi-granulation model are discussed. Moreover, the certainty factor of the decision rule and acquired method from point of multi-granulation are then given.(3) The lower approximation of existing rough set model will include few objects in dealing with inconsistent intuitionistic fuzzy decision system and affect the acquisition of deterministic decision rules. Variable consistency rough set model of inconsistent intuitionistic fuzzy decision system based on the consistency degree is proposed in this paper. Part of the objects being missed can be admitted to the lower approximations controlled by an index called consistency level. Some important properities are then analyzed. The classification quality under the consistent level as well as the reduction method and the acquisition method of optimal decision rules are then given.(4) Under the background of traffic accident historical statistic data, the limitation of existing single decision attribute rough set model in dealing with traffic accident cause is analyzed. A group decision attributes rough set model is proposed in this paper and the properities of the model as well as its relationship with single decision attribute model are then analyzed. Based on the two indice, namely hidden danger index(condition attribute) and the accident index (decision attribute) , the road traffic accident factor correlation analysis and decision rules extraction are researched which provide reference for traffic management decision-making.