vai al contenuto della pagina vai al menu di navigazione

Statistics Seminar: Bivariate Teissier Model and Applications

15/09/2017 dalle 14:30 alle 17:10

Dove Dipartimento di Scienze Statistiche, Aula Seminari - I piano

Aggiungi l'evento al calendario

Nikolai Kolev
University of Sao Paulo (Brazil)

The baseline model under consideration has been first introduced by the French biologist Teissier (1934) considering a mortality of several domestic animal species protected from accidents and disease, i.e., dying as a result of a "pure aging". Specifically, for a non-negative random variable X, the model is defined by the survival function

P(X > x) = exp{x + 1- exp(x)}, x≥ 0:

It is direct to check that the corresponding mean residual life function E[X-x|X> x] = exp{-x}. Teissier's distribution is motivated by the empirical fact that many vital functions are decaying exponentially.
We first give historical remarks about the forgotten univariate Teissier's model. We introduce a bivariate version of the Teissier distribution and outline its basic properties. The corresponding copula is obtained and applications are discussed.

Reference: Teissier, G. (1934). Recherches sur le vieillissement et sur les lois de mortalite. Annales de Physiologie et de Physicochimie Biologique 10, 237-284.

Sabrina Mulinacci