Statistical Analysis Of ALT Under Mixed Grouped Data

Compared with complete life testing, truncated life testing are more time-saved, but we must track the whole testing and record the products' invalid time, so lots of manpower and material resources are needed in those truncated life testing. When we record the number of invalid products in the fixed time intervals but the products' invalid time in life testing, we will save lots of time, manpower, and expenditure. But this will increase more difficulties in our statistical analysis. So we discuss this kind of ALT: in the low stresses, we get grouped data; in the high stresses, we get truncated data.This paper mainly discusses EM Algorithm's applications in estimating the parameters of accelerated model and present the estimator of average life, when the products' life distribution is exponential distribution and the data is mixed grouped data. We will focus on constant-stress models and step-double-stress models.