Copula-Robust Cramér–Rao Lower Bound underDependent Random Censoring
DOI:
https://doi.org/10.56143/ujmcs.v2i3s.3Keywords:
survival analysis; Cramér–Rao Lower Bound; Fisher infomation; copula function; robust Godambe information; Censoring.Abstract
In survival analysis and reliability theory, the classical Cramér–Rao lower bound assumes independence between lifetime and censoring variables. This paper extends the bound to dependent censoring using a copula-based likelihood. We derive a robust Godambe information bound and prove a generalized Cramér–Rao inequality. The result reduces to the classical bound under independence and shows increased variance bounds with stronger dependence.
References
[1] Abdushukurov A. A.: Extension of relative-risk power estimator under dependent random censored data. CMFD 68 (1), 1-13, (2022).
[2] Godambe V. P.: Estimating Functions. Oxford University Press, (1991).
[3] Abdushukurov A. A., Muradov R. S.: On estimation of conditional distribution function under dependent random right censored data. Journal of Siberian Federal University. 7 (4), 409-416, (2014).
[4] Nelsen R. B.: An Introduction to Copulas.Springer. New York, (2006).
[5] Abdushukurov A. A.: On Fisher-Bhattacharyya Information Function and Lower Bounds under Right Random Censoring. Lobachevskii Journal of Mathematics. 45 (9), 1-10, (2024). DOI: 10.1134/S1995080224603011
[6] Abdushukurov A. A., Muradov R. S.: On estimation of conditional distribution function under dependent random right censored data. Journal of Siberian Federal University. 7 (4), 409-416, (2014).
[7] Abdushukurov A. A., Muradov R. S.: Copula-Based Fisher Information and Cramér–Rao Lower Bound under Dependent Censoring. Abstracts of International Conference on "Current Problems of Mathematical-Physical Equations and Information Technologies. Part 1, 298-300, (2026).
