An Improved Multi-modal Data Decision Fusion Method Based on DS Evidence Theory

Abstract

The method from DS evidence theory based multi-modal information decision fusion uses the classification structure information which the correct and error classification information provided by the classifiers. These two types of information affect the fusion results of DS evidence theory. This paper proposes a new method(DShW) for correct and error classification information in the balanced classification structure information based on DS evidence theory. That is, a method based on inertia weight normalization is introduced in the confusion matrix. To adjust the specific gravity of correct and error classification in classification structure information by changing the size of the value h, so as to achieve the purpose of balancing correct and error classification information. By comparing with other classifiers, we find that the DShW method effectively improves the accuracy of decision fusion.