Answers are grounded in this paper and cite their sources.
Assesses independence testing in models, suggesting effective procedures using rank-based measures.
Generalized partially linear model is a semiparametric model of interest that extends the generalized linear model incorporating a nonparametric component. This article delineates testing procedures on the independence of regressors and error component in a generalized partially linear model based on rank-based nonparametric measures of association—Spearman’s ρs, Kendall (1938)’s τ and Bergsma and Dassios’s τ∗. Such tests of independence are of utmost interest, since assumption of independence among regressors and error component in effect yields a reliable estimate. Moreover, using the prescribed measures of association, consistent tests are constructed by defining V-statistics based on the paired values on a general order difference of response and the same ordered difference of estimated response. The assessment of the performance of the test statistics is based on their asymptotic powers against a sequence of contiguous alternatives. The asymptotic distributions of the proposed test statistics are deduced under both the null and contiguous alternative hypotheses. Finally, the efficacy of the proposed measures is illustrated through considering a real-life data example.
No takes yet. Share an insight, caveat, or question.
Das et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: