A Research Into The Evidence Reasoning Theory Of Two-Dimension Progressive Framework And Its Application

With the advancement of economic society, science and technology, nowadays, people have encountered increasingly complex decision problems and constantly rising decision associated information contents. Owing to the evidence theory exhibits its special advantages, evidence decision flow accords well with the cognitive process of human beings. This provides people with an effective decision method. However, traditional evidence theories are constructed based on one dimensional evidence recognition theory. They merely reflect the reliability information determined by evidence source and fail to present the characteristic information of evidence source itself or the information for the process of determining reliability. This greatly influences the processes in decision such as evidence processing and combination. Therefore, authors aim to use a two-dimension evidence framework to reflect the process and characteristic information of evidence source based on traditional evidence framework. Then, the evidences processed and combined by two-dimension information are modified. This research is of theoretical and practical significance in extending theoretical connotation and meeting the requirements of efficient and accurate decision performances by fully using precise evidence information.This research takes the evidence reasoning theory of two-dimension progressive evidence framework as a research main objective and the evidence reasoning of two-dimension progressive recognition framework as mainline. Then it proposes the principle and general flow of constructing two-dimension progressive evidence recognition framework by utilizing the mechanism which applies two-dimension evidence framework through using extension theory and cooperative analysis. Moreover, the authors investigate the problems regarding evidence combination of the two-dimension progressive evidence framework and put forward an evidence packet combination method so as to fully fit two-dimension information into traditional evidence frameworks. Furthermore, an information fusion model for stock rating of securities analyst is established based on the two-dimension progressive recognition framework so as to be applied in information fusion of stock rating of securities analyst. The research content and highlights for this paper are presented as follows:1. This research proposes a two-dimension evidence reasoning theory of two-dimension progressive framework to deal with decision problems in detail.Based on the traditional evidence frameworks, the authors put forwards a two-dimension information framework of modifying and illustrating evidences. Besides, the mechanism of constructing two-dimension progressive evidence framework on the basis of extension theory and collaborative analysis is revealed. The framework proposed is able to store more evidence information than those traditional ones. Furthermore, this research gives two ideas for direct and indirect combinations of evidence fusion using two-dimension progressive evidence framework respectively. In this work, the idea for direct evidence combination is used. In addition, the principle and general flow of constructing two-dimension progressive recognition framework to cope with the decision problems are demonstrated. The empirical analysis of two-dimension recognition framework for progressive evidence is carried out by taking the peer review initiated by the committee of National Natural Science Foundation of China.2. This works provides an evidence packet combination method based on two-dimension progressive evidence framework.This study analyzes the general flow of the evidence packet combination method based on two-dimension information. It proposes an evidence classification method on the basis of the two-dimension framework. This works demonstrates general principle for evidence packet:the evidence source is divided into several packets or types; Besides, the evidence combination method of each packet and the problems resulting from evidence combination in different packets are explored. According to the research, the key of the fusion of the evidences among different packets is to construct a mapping relationship between the two-dimension evidence information and evidence modifying factors. Lastly, in the research of evidence decision rule, the authors study the ways of determining and selecting decision schemes based on evidence fusion results. The research also investigates a prospect theory based idea for evidence decision.3. An information fusion model for stock rating of securities analyst is established based on the two-dimension progressive recognition framework.This study analyzes the shortcomings of existing statistics based integrative information models of securities analyst. Aiming at the scenario of information fusion of stock rating for securities analyst, the securities analysts’ information concerning stock rating and earnings forecast are used to construct a two-dimension progressive evidence framework. Moreover, the stock rating information from many securities analysts are fused into one item of evidence information using a two-dimension framework information based evidence packet combination method. Then, the data obtained from Guotaian database are applied to conduct empirical research. The comparative analysis of the method proposed and traditional methods are performed in terms of correctness and differentiation.