CS-PALT Combined With Different Censoring Techniques On Gompertz Distribution: Some Inferences

Abstract

The major objective of the present article is to study the effect on Bayes risks for different censoring techniques combined with constant-stress partially accelerated life test. The underlying distribution of this study was considered as two-parameter Gompertz distribution. The Bayes risk under invariant LINEX loss function has been obtained for the parameters under study. The censoring techniques have been used here, Type-I progressive hybrid, Progressive Type-II, Type-I progressive and Type-II censoring. Numerical illustration based on real and simulated data has been carried out. The Metropolis-Hastings algorithm has also been combined with the simulation study for improvement in precision of inferences.

Key words: Constant-Stress Partially Accelerated Life Test (CS-PALT), Type-I Progressive Hybrid (T-IPH) Censoring, Progressive Type-II (PT-II) Censoring, Type-I Progressive (T-IP) Censoring, Type-II (T-II) Censoring; Invariant LINEX Loss function (ILLF).

L’objectif principal du present article est d’ ´ etudier l’effet de diff ´ erentes ´ techniques de censure sur les risques bayesiens lorsqu’elles sont combin ´ ees dans ´ un test de survie partiellement accel´ er´ e sous contrainte constante. La distribution ´ sous-jacente de cette etude a ´ et´ e consid ´ er´ ee comme une distribution de Gompertz ´ a deux param ` etres. Le risque Bayes la sous fonction de perte LINEX invariante a `
et´ e determim ´ e pour les param ´ etres ` etudi ´ es. Les techniques de censure utilis ´ ees ´ ici sontM: la hybridation progressive de type I, la censure progressive de type II, la censure progressive de type I et la censure de type II. Une illustration numerique bas ´ ee sur des donn ´ ees r ´ eelles et simul ´ ees a ´ et´ e r ´ ealis ´ ee. L’algorithme ´ de Metropolis-Hastings a egalement est utilis ´ e dans l’ ´ etude de simulation pour ´
l’amelioration de la pr ´ ecision des inf ´ erences.