Reliability Evaluating Theory for Data Sample with Unknown Priori Information

Abstract

Hierarchical maximum entropy Bayesian method is proposed to establish the reliability evaluation model for data sample with unknown priori information. The probability density functions are calculated for different time series using the maximum entropy principles and the Bayesian method is utilized to obtain the posterior sample information for different time intervals. The values of performance continuity relative reliability for time series data sample can be calculated during the corresponding time intervals and the failure degree maintaining optimum performance status in the future can be predicted.