A Field Data Processing Approach Based on Grey Entropy

Traditional statistical data processing approach needs to know the distribution regularity of samples. But in filed test, the sample distribution regularity usually can´t be known or has many likelihoods due to complex external environments. The traditional statistical approach will give an assumptive model about the sample distribution for parameter estimation. The assumption usually takes new error to the parameter estimation value, and makes the reliability of the parameter estimation approach lower. This paper proposes a new data processing approach based on grey entropy, from the view of the topology of the sample space and the distances between samples. The definitions of grey distance measure and grey entropy are given. The methods of removing gross error and parameter estimation based on the grey entropy definition are proposed in the paper. Finally, the simulation results show that this approach is feasible.