The Dirichlet process is perhaps the most popular prior used in the nonparametric Bayesian inference. This prior which is placed on the space of probability distributions has conjugacy property and asymptotic consistency. In this thesis, our concentration is on applying this nonparametric Bayesian inference on the Tukey depth and Tukey median. Due to the complexity of the distribution of Tukey median, we use this nonparametric Bayesian inference, namely the Lo’s bootstrap, to approximate the distribution of the Tukey median. We also compare our results with the Efron’s bootstrap and Rubin’s bootstrap. Furthermore, the existing asymptotic theory for the Tukey median is reviewed. Based on these existing results, we conjecture that the bootstrap sample Tukey median converges to the same asymp- totic distribution and our simulation supports the conjecture that the asymptotic consistency holds.

Dado un conjunto de datos, o una distribución, en un espacio de dimensión mayor a uno, las proyecciones aleatorias consisten en proyectar los datos, o calcular la marginal de la distribución, en un subespacio de menor dimensión que ha sido elegido de forma aleatoria. En nuestro caso de dimensión uno. En esta tesis presentamos dos aplicaciones de las proyecciones aleatorias. La primera es una definición de profundidad, que es computacionalmente efectiva, aproxima a la conocida profundidad de Tukey y es válida tanto en espacios multidimensionales como funcionales. La segunda es un test de Gaussianidad para procesos estrictamente estacionarios, que rechaza procesos no Gaussianos con marginal unidimensional Gaussiana. / A random projection consists in projecting a given data set, or in computing the marginal of a distribution, on a randomly chosen lower dimensional subspace. In our case, it is of dimension one.In this thesis, we present two applications of the random projections. The first one is a new definition of depth that is computationally effective, approximates the well-known Tukey depth and works as much in multidimensional spaces as in functional. The second is a test of Gaussianity for strictly stationary processes, which rejects non-Gaussian processes with Gaussian one-dimensional marginal.