Answers are grounded in this paper and cite their sources.
First, we examine the finiteness of a commonly used formulation of cumulative Tsallis entropy and show that, for distributions with unbounded support, the defining integral is generally divergent. This motivates an alternative formulation based on the cumulative distribution function (CDF) that is finite under appropriate moment conditions and admits a natural interpretation in terms of the mean of a generalized maximum. Second, we develop a comprehensive framework for quantifying uncertainty in concomitants of k-record values Z[n,k] arising from the Sarmanov (SAR(ζ)) family of bivariate distributions, using cumulative Tsallis entropy (CTE) as the primary information measure. We derive explicit closed-form expressions for the CTE, its past-lifetime dynamic variant, and the associated Rényi entropy for the concomitants Z[n,k] under the SAR(ζ) dependence structure. A quantile-based reformulation is employed throughout, enabling the results to hold for arbitrary marginal distributions without requiring closed-form CDFs. Furthermore, we establish several fundamental theoretical properties, including uniqueness of the quantile density characterization, relationships to the mean inactivity time, and monotonicity classes that provide nonparametric criteria for model discrimination. To facilitate practical implementation, we introduce a nonparametric kernel-based estimator for the Tsallis entropy, with comprehensive comparisons of bandwidth selection methods. Finally, a real-world dataset is used to illustrate the proposed methodology.
No takes yet. Share an insight, caveat, or question.
mansour et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: