We propose a bivariate exponential model with exponential marginal densities, correlated via a latent random variables and with finite probability of simultaneous occurrence. We extend this model to a bivariate Erlang type distributions with same shape parameter. We propose an estimation method and illustrate our results with real data.
The 4th International Conference on Neural, Parallel & Scientific Computations (ICNPSC 2010). Proceedings of Neural, Parallel, and Scientific Computations, Volume 4, 2010 (Atlanta, GA 11-14 August, 2010) p. 161-168