Incomplete information reasoning method with mode switching mechanism

Abstract

The invention provides an incomplete information reasoning method with a mode switching mechanism. The incomplete information reasoning method comprises the following steps: selecting effective attribute information, determining a decision variable, and constructing a modular switching reasoning evidence network model; determining an identification framework, and initializing a network model; calculating and initializing a reliability distribution function among the node variables according to the logic relation of the node variables; carrying out completeness degree judgment on the input evidence at the current reasoning moment, selecting an input mode in combination with an evidence type, carrying out evidence reliability function correction according to the input mode, and adjusting the network model; inputting the correction evidence, adaptively adjusting node weights and updating a reliability distribution function between the nodes; carrying out network reasoning by using a three-step space-time fusion mechanism to obtain probability distribution of decision variables under an identification framework of the decision variables; a normalized mathematical expectation of the decision variable is calculated. In the face of a complex decision problem under mixed incomplete information, a continuous and accurate reasoning result can be given, and the autonomy and intelligence of a decision system are comprehensively improved.

Description

technical field [0001] The invention relates to the technical field of multi-source heterogeneous information fusion, in particular to an incomplete information inference method with a mode switching mechanism. Background technique [0002] Information reasoning in reality is mostly a multi-source heterogeneous information fusion problem in an uncertain environment. It not only needs to consider how to express and fuse entities represented by a variety of quantitative and qualitative data and the complex connections between them, but also needs to be able to Effectively deal with random data defects (distortion, error or even missing, etc.) that may appear in the continuous reasoning process. Although the traditional method can solve the problem of uncertain information reasoning in the general sense, it generally does not consider the effective processing and fusion of information in the case of mixed various incomplete data, resulting in the inability to give continuous co...

AI Technical Summary

Problems solved by technology

Although the traditional method can solve the problem of uncertainty information reasoning in the general sense, it generally does not consider the effective processing and fusion of information in the case of mixing multiple incomplete data. reasoning about outcomes, thereby adversely affecting subsequent decisions

Images