Research On Three-way Decision Model And Its Application Based On Probabilistic Graph

This dissertation comes from the key scientific and technological project of Education Department of Henan Province, China(No.14A520082) “Research on Constructing of the Intuitionistic Fuzzy Rough Implicators and Its Application for Stable Control System”, Foundation and Advanced Technology Research Program of Henan Province, China(No.132300410174) “Research on Interval-valued Fuzzy Implicators and it’s Algebraic Structure”, and the key scientific and technological project of Xinxiang City, China(No.ZG14020) “Research on Constructing of the Intuitionistic Fuzzy Rough Implicators and it’s Application for Decision and Control”.As a combination of probability theory and graph theory, probabilistic graphical model provides a natural and intuitive solution for uncertain and complexity problems in the field of intelligent information processing. With the improvement of the probability graph learning theory and reasoning algorithm, the probability graph representation of the complex knowledge is widely used in risk assessment, medical diagnosis, pattern recognition and other fields, through the establishment of accurate model of big data and realizes the reasoning. Three-way decision as a decision model of human cognition, including positive region, boundary region and negative region, represents making a decision of acceptance, rejection and non-committed respectively. combining the probabilistic graph with the three-way decision, the intelligent information processing has become a research hotspot in various fields, but there is few research about the the view of inference to determine the conditional probability Based on the original reasoning method, the problem of probability can expand into the decision fields, by combining the three regions of the three-way decision. It provides a more comprehensive and reasonable solution for dealing with incomplete information. In this paper, the main research works are as follows:(1) The uncertain information reasoning problems, stretch out in all directions in the field of decision-making, through three decision based probabilistic inference in the domain of positive and negative, domain and edge boundaries, and puts forward the three decision-making model based on precise reasoning, and verified by practical teaching evaluation for the effectiveness of the model.(2) On the basis of general reasoning and decision making when considering the cost function for the fuzzy number model, combined with probability graph, fuzzy sets and three decision theory, is proposed based on fuzzy probability graph and three decision-making model, using the membership function instead of the specific numerical cost function, early enough new three decision-making model, and use of stock risk assessment example verifies the feasibility of the model.(3) For target recognition, the traditional probabilistic graphical model, extended to three decision model, based on the minimum risk decision model of evidence reasoning and three supporting decision making, presents evidence reasoning model and the probability map of three based on the decision, through the simulation example of sea target recognition, to illustrate the effectiveness of the model.Through probability graph and three decision-making under the general model, the fuzzy relation of the probability of the figure with three decision-making model and based on the probability map with three decision-making evidence reasoning model research, enrich and perfect the probability graph with three decision theory, and expand the application range of probability diagram and the three decisions.