Numerical differences in the survival probabilities obtained by the cumulative‐hazard and kaplan–meier estimators of the reliability function

摘要

AbstractThe paper examines two very important non‐parametric estimators of the reliability function which are frequently used in reliability practice to estimate survival probabilities. The paper proves that when both of them are calculated for the same data set, the survival probabilities obtained using the cumulative hazard approach are larger in magnitude than the ones given by the method proposed by Kaplan and Meier, and provides the mathematical formula quantifying the magnitude of this differenc