Research For Three-Way Rules Based On Decision Preference In Incomplete Information Systems

The knowledge system is called incomplete information system when one or more attribute values are missing due to unavoidable external factors such as the limitation of information acquisition,data omission and noise.Three-way decision is a powerful tool to deal with incomplete information system.Three-way decision divides those objects with insufficient or inadequate decision information into the boundary domain,meaning that the objects need further investigation,as the third decision choice between accepting and rejecting decision making.However,among the existing three-way decision models based on incomplete information systems,there are few comprehensive strategies that consider both the local importance of attribute values in information system and the subjective initiative of decision makers in the decision-making process.When faced with a decision,people’s tendency to choose one of several alternatives is called decision preference.Decision preference is a research hotspot of three-way decision making.This paper studies the three-way decision rules based on incomplete information system from subjective and objective aspects as follows:(1)This paper discusses decision preferences of decision makers for the importance of attribute values in different classes from an objective perspective.Firstly,a new weight construction method is defined according to the degree of local contribution of conditional attributes in the classification process,and an improved calculation method for similar classes is proposed.The effectiveness of this method is proved by comparing it with the traditional 0-1 matching difference method through an example.Then,a weighted three-way decision model based on specific class attributes in incomplete information systems is constructed.This method is effective for processing data sets with few conditional attributes.Finally,two real cases in the UCI database are used to compare the proposed model with the three-way decision model based on incomplete information system.Experiments show that the proposed algorithm is superior to the latter in classification results.(2)In order to reflect the subjective preferences of decision-makers in the decisionmaking process,the prospect value function is constructed under the cumulative prospect theory,which can reflect the difference of interval value with different interval radius and expectation at the same time.Furthermore,some properties of prospect value function and their proofs are given.By introducing the cumulative prospect theory into the three-way decision model based on incomplete information,the threshold value calculation method based on the decision prospect is presented.By discussing the calculation methods of loss function and gain function under uniform distribution,we judge the loss and gain relationship between two interval value when there is inclusion relationship.Finally,in order to verify the effectiveness and feasibility of proposed algorithm,the prospect value for decision making and threshold changes are analyzed under different risk attitudes and different radius of the interval valued.In addition,our model is compared with the interval-valued decision rough set model,and the experimental results show that the former has better decision prospect.