For the bivariate posterior densities, two methods are supported. The default is based on fast Fourier transforms with a Gaussian kernel and taken from the Kernel Density Estimation Toolbox of Christan Beardah. The second is based on linear diffusion processes and uses Zdravko Botev's code kde2d, which can be downloaded from Mathworks website and which is discussed in Botev, Grotowski and Kroese (2010). YADA can create both surface and contour plots of user selected pairs of bivariate densities.
Page url: http://www.texlips.net/yada/help/index.html?bivariate_posterior_densities.htm