Improvement and application of an uncertainty management method based evidence theory

摘要

Aiming at earthquake prediction field, this paper puts forward a multi-expert multi-attribute uncertainty management algorithm based on DS evidence theory. Seismic signal has so much uncertainty that traditional mathematical statistics method cannot solve. Considering DS theory is an excellent uncertainty reasoning method, we extract three instantaneous attributes of plural seismic signal as prediction characteristics, and simultaneously modify DS theory to fusion multi-attribute characteristics by the introduction of Jffreys distance function. Experimental results and analyses demonstrate that, through the proposed multi-expert multi-attribute uncertainty management algorithm, we can clearly describe geological phenomenon and sedimentary processing. Thus, it has great application value.