The Application Of The Principle Of Maximum Entropy And Bayesian Methods In The Measurement Data Processing

No measurement, No science. Modern natural science mostly can not depart from measurement. In measurement field, the measurement process divides into two parts: measurement data obtaining and measurement data processing, both of them play important part in measurement result obtaining. The mainly content of this paper is the new theory of measurement data processing. The applications of Maximum Entropy Method (MEM) and Bayes in data processing are researched in this paper.The MEM applications in this paper include confirming whether the distribution of the data according with a certain distribution and obtaining the probability distribution function through the measurement data.The Bayes applications in this paper include the different prior distributions'influence upon the result and two methods of getting likelihood function: step by step amending and group amending and their differences. Instructional conclusions are drawn through the research. The Bayes theory is also used to get two familiar distributions: Normal and Poisson.An innovational combination is made in this paper, which is the combination of MEM and Bayes. This method can use both the two methods'advantages to process the data, and can get better results.In the last part of this paper, an experiment was held to get the measurement data used in data processing. The measurement data were used to validate the conclusions.