A New Lifetime Distribution: Exponentiated Exponential-Pareto-HalfNormal Mixture Model for Biomedical Applications

TL;DR

The paper introduces a novel hybrid distribution model, EEPHND, combining multiple distributions to better fit complex biomedical survival data, with demonstrated superior performance on lung cancer data.

Contribution

It develops the EEPHND model, providing closed-form expressions and demonstrating its effectiveness in modeling biomedical survival data.

Findings

Outperforms competing models in goodness-of-fit

Achieves a Concordance Index of 0.9997 on lung cancer data

Captures both early-time symmetry and long-tail behavior

Abstract

This study introduces the Exponentiated-Exponential-Pareto-Half Normal Mixture Distribution (EEPHND), a novel hybrid model developed to overcome the limitations of classical distributions in modeling complex real-world data. By compounding the Exponentiated-Exponential-Pareto (EEP) and Half-Normal distributions through a mixture mechanism, EEPHND effectively captures both early-time symmetry and long-tail behavior, features which are commonly observed in survival and reliability data. The model offers closed-form expressions for its probability density, cumulative distribution, survival and hazard functions, moments, and reliability metrics, ensuring analytical traceability and interpretability in the presence of censoring and heterogeneous risk dynamics. When applied to a real-world lung cancer dataset, EEPHND outperformed competing models in both goodness-of-fit and predictive accuracy, achieving a Concordance Index (CI) of 0.9997. These results highlight its potential as a flexible and powerful tool for survival analysis and biomedical engineering.