Research Of Flow Network And Decision Algorithm Based On Rough Set

Inductive learning and classification of object from data set are very important areas in artificial intelligence, in order to discover hidden, unknown, potentially useful knowledge in the data, and to search the rule and the general mode from the large data set in essence. In recent years, a lot of theories based on inductive learning have been researched, and a number of techniques are developed to deal with imprecise data. The most successful technique is rough set theory, which is a powerful tool about data reasoning. Rough set theory has been successfully used in machine learning, knowledge acquisition, pattern recognition and decision support systems, etc.Rough set theory that was put forward by Pole Z. Pawlak in 1982 is a new data analysis theory of analyzing and dealing with uncertain and incomplete data. It makes use of the equivalence relations to measure the indetermination degree of knowledge and it doesn't need any knowledge outside of the data which needs to be processed. Therefore the error caused by subjective appraisal can be avoided. So the studies of decision-making based on rough set theory have widespread application prospect.Reduction of attributes and reduction of attributes'value are the base of mining decision rules from decision table using rough set theory, and mining decision rules is one of the most important research fields in rough set theory. Based on the important degree of condition attributes and discernible matrix, an effective condition attributes reduction is obtained. A notation of inconsistency of decision table is given, the attributes'value reduction is got by the important degree of attributes and decision rules are obtained in the premise of maintaining the inconsistency of the decision table. Using this method, the set of decision rules which are mined from decision table satisfies the properties of independent, covering universe, admissible and consistency, that is, it is a decision algorithm in the decision table. And, rough set flow network is studied in this paper, which combines rough set flow network, decision algorithm and Bayes'theorem. Lastly, a model of risk rule mining based on rough set and Bayes'theorem is given.The works done in the paper are as follows:(1) The elementary knowledge of rough set theory is studied, several core concepts of rough set theory are introduced, and the concept of the important degree of condition attribute and the notation of inconsistency of decision table are given in this paper, which have laid the foundation for the reduction of attributes and the mining of decision algorithm in the later.(2) The properties of decision algorithm are discussed in detail in this paper. Through the research, we find that decision algorithm satisfies the total probability theorem and the Bayes'theorem. By using the Bayes'theorem based on rough set, we only need to calculate the strength of the decision rule, and then calculate the certainty factor and the coverage factor of the decision rule. The computation complexity is simplified greatly.(3) Reduce attribute by using the important degree of condition attribute in the discernible matrix. A method of rough set based decision algorithm mining is given in the premise of not changing the inconsistency of decision table. The algorithm which is mined in this method is the smallest set of the decision rules which cover all objects. This method is also a very important innovation of this paper.(4) Rough set flow network which is a new mathematical model of decision processes is studied. The whole network represents a decision algorithm. And the properties of flow network are researched. The concepts of strength, certainty and coverage factors are extended to the path and connection.(5) The data is reduced by rough set and the data which is reduced is trained by the Bayes'theorem. A model of risk rule mining based on rough set and Bayes'theorem is given. The model combines rough set and Bayes'theorem and is applied in the risk management of the IT project.Many problems in the decision analysis by rough set theory need to be discussed. And there are many problems in the study of this paper. The according work will be done in the future.