The thesis contributes to the literature on Bayesian nonparametrics by proposing two approaches, the first one related to time series analysis with a focus on sparsity of the matrix of coefficients and the second one to ...