Random Forests in combination with Stability Selection allow to estimate stable conditional independence graphs with an error control mechanism for false positive selection. This approach is applicable to graphs containing both continuous and discrete variables at the same time. Its performance...