A1391
Title: Least-squares estimator-based inference for survival and reliability data
Authors: Bowen Liu - University of Missouri - Kansas City (United States) [presenting]
Malwane Ananda - University of Nevada Las Vegas (United States)
Sam Weerahandi - X-Techniques (United States)
Abstract: Statistical inference for survival and reliability measures is essential in engineering, biomedicine, and risk analysis. An inferential framework based on least-squares estimators (LSEs) for survival probabilities, reliability functions, and related quantities in location-scale families is proposed. Using these estimators, statistical inferential procedures are developed and bootstrap methods for interval estimation are investigated. Results show that LSE-based confidence intervals perform particularly well compared with likelihood-based methods. In simulation studies, they achieve strong coverage probabilities and favorable interval efficiency, often outperforming likelihood-based methods, especially in small samples where asymptotic approximations may be unreliable. The proposed methodology applies to several widely used lifetime models, including the Weibull, log-logistic, and Gumbel distributions. Real-data examples further illustrate the practical effectiveness of the proposed methods in survival and reliability analysis.
