误差环境中参数识辨前测量信息的熵描述

Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2018-04-17 Cooperative journals: 《计算机应用研究》

李静 丁海洋 任学尧

Abstract： In the parameter identification of inverse problems, it was well-known that the information quantity contained in measurement information influences the reconstruction precision of parameter directly. How to describe the information quantity in measurement information plays an important role in the selection of the number of measurement points. According to the Bayesian method, the probability of the parameter sample to be identified is calculated by combining the priori probability reflecting prior information and the likelihood probability reflecting measurement information. The evaluation factors are calculated by the maximum entropy and information entropy of the parameter. This paper introduced the information entropy to describe the information quantity of measurement information with error before parameter identification. Numerical tests showed that this computing method proposed could describe the measured information with error effectively and determine the measured information in practice.

 Hits 1908  Downloads 1025  Comment