Entropy-based characterization and Bayesian hierarchical modeling of the new X-Lindley distribution
摘要
This study introduces and systematically analyzes the New X-Lindley Distribution (NXLD), a one-parameter lifetime model extending the classical Lindley family. We derive closed-form expressions and simulation-based evaluations of various entropy measures, including Shannon, Rényi, and Tsallis entropies, and information divergence criteria such as the Kullback–Leibler and Jensen–Shannon divergences. Emphasis is placed on the distribution’s potential for modeling uncertainty and tail behavior in real-world stochastic systems. Additionally, a Bayesian hierarchical framework is proposed to capture latent variability in the rate parameter across spatiotemporal domains. Application to astrophysical data, specifically, photon interarrival times from pulsars and gamma-ray bursts, demonstrates NXLD’s superior fit over conventional models. The main contribution of this work lies in developing the NXLD entropy framework, which provides a more robust and flexible representation of complex astronomical signals compared to traditional entropy measures. This work positions NXLD as a competitive tool for modeling heavy-tailed and non-stationary data in astronomy and other scientific domains.