A FUZZY APPROACH TO IMPROVE SURVIVAL ESTIMATES: REVISITING THE BRESLOW’S ESTIMATOR

Read full text Summarise with VIRAT
Authors: R Jaisankar, M Siva
Cited by: 0
References: 6

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Censoring is a natural aspect of survival analysis, which occurs when the actual time when an event occurs is not observed completely or when the individuals fail to experience the event during the study interval. To handle this type of incomplete information, non-parametric methods are frequently used, with the estimator by Breslow being one of the most commonly used methods of estimating the cumulative baseline hazard in the Cox proportional hazards model. Although widely used, the estimator can lose accuracy in the case where survival times are unknown, approximate or are captured ambiguouslycircumstances that are likely to prevail in real-life medical data sets. This paper suggests a more refined form of the estimator introduced by Breslow, which incorporates fuzzy membership functions to better reflect imprecise data of time-to-event information. The fuzzy extension provides a more abundant and adaptable approach to defining uncertainty than the traditional crisp definitions by admitting partial degrees of membership to time intervals. This framework is especially useful in clinical studies, where patient-reported timings, non-adherence to follow-up, and missing records may add a degree of vagueness that is impossible to circumvent. The applied fuzzified estimator will enhance the precision and stability of the baseline hazard estimator, which is a more versatile tool in the analysis of complex survival data and will lead to more credible inferences in medical decision-making.

Fields of Study & Subjects

Fuzzy Systems and Optimization Probability and Statistical Research Statistical Distribution Estimation and Applications

Publication details

Journal / source
Reliability: Theory & Applications
Publication date
1 January 2026
Language
EN