The Bayesian Estimation Of The Relationship Between Chip Aging And Voltage

LED chip is the best choice to make signs and bright words because of long life,low cost and high brightness.No matter what daily lighting,light indicator,or decoration,the LED chip focus on efficiency and stability.Some company did a research that relation of the specific value and voltage where the value represent to light strength of long-worked chip and initial strength of this chip.We found that the traditional normal linear model whose error terms is set by Gauss-markov theory was no longer applied into the research.In this paper,according to the actual data,this paper comes up to a linear model whose error term obeys Beta distribution.This paper uses Bayesian and Empirical Bayesian to estimate parameters with Gibbs sampling and M-H algorithm.And this model was applied to the actual data about chip and voltage.It forecasts the voltage when the loss of chip is zero after long time working.This is of great significance to adjust the working voltage of LED chip properly and ease its aging.