Evaluation for Data Fusion System Based on Generalized Relative Entropy
WU Xiu-qing · Jisuanji fangzhen · 2006
Evaluation for data fusion system is discussed based on relative entropy.The method measures relative information gain with generalized relative entropy.Some of the most common types of uncertain information can be represented in the framework of fuzzy measure.Shapley entropy is introduced to the framework to calculate the amount of information in a fuzzy measure.The relative entropy to fuzzy measures is generalized.Normalized relative entropy is utilized to evaluate data fusion system.This method measures the information gain of data fusion system in the framework of fuzzy measures.Its robustness is good.It is a effective method to evaluate data fusion system.