Integral estimate method for incomplete data

A new method of integral estimate for incomplete data is presented in this paper. The integral parameter estimators and their covariance are derived. The integral estimate method for the location-scale distributions is discussed at some length. The method extends traditional regression analysis that is only suitable for complete data to incomplete data. Thus it is applicable even when there is only one failure data while the traditional method requires more failure data in a state. As the crosswise information among the incomplete data in different states is exploited and applied, the number of specimens for determining the distribution parameters in all states by the method is about the same as that of the specimens in one state by the traditional method when the precision is equal, which shows the method has the advantage of small sample.