A new finite approximation for the NGG mixture model: An application to density estimation

A new class of random probability measures, approximating the well-known normalized generalized gamma (NGG) process, is defined. The new process is built from the representation of the NGG process as a discrete measure, where the weights are obtained by normalization of points of a Poisson process larger than a threshold ?. Consequently, the new process has an as surely finite number of location points. This process is then considered as the mixing measure in a mixture model for density estimation; we apply it to the popular Galaxy dataset. Moreover, we perform some robustness analysis to investigate the effect of the choice of the hyperparameters.