Covariate significance testing in the conditional bivariate distribution function

Abstract

One significant goal in recurrent events analysis is the estimation of the bivariate distribution function. Estimating this function for censored gap times is crucial across various fields and applications, as it helps elucidate recurring events and their underlying patterns. ’Gap time’ refers to the duration between successive occurrences of an event, while the bivariate distribution function describes the joint probability distribution of two such gap times. Despite considerable progress in this area, most ap-proaches do not account for the influence of covariates. In this paper, we introduce a viable nonparametric method for estimating the bivariate distribution function conditioned on current or past covariate measures. In addition to this, the primary aim of this pa-per, however, is to introduce ideas of possible methods for testing the significance of covariates in the estimation of the conditional bivariate distribution function.