An improved approach for incomplete information modeling in the evidence theory and its application in classification

Incomplete information modeling and fusion under uncertain circumstances remain a significant open problem in practical engineering. In this study, the Dempster–Shafer evidence theory is extended to the generalized evidence theory, and the above problem is addressed from the perspective of open-world assumptions. An improved method is proposed to model incomplete information, where the generalized basic probability assignment (GBPA) is generated using the Gaussian distribution model. First, we constructed the Gaussian distribution based on the mean and variance calculated from the training set. Then, we modeled the potential incomplete information with the GBPA of the empty set by matching the test sample with the constructed Gaussian distribution model. Next, we identified and recognized the unknown object by fusing the data with the generalized combination rule. Finally, classification experiments and comparative studies were conducted to illustrate the superiority and effectiveness of the proposed method.