| Popis: |
This study evaluates both a variety of existing causal inference methods and a variety of ensemble methods. We show that: (i) individual causal network methods vary in their performance across different datasets, so a method that works poorly on one dataset may work well on another; (ii) a Bayesian ensemble method leads overall to better results than using the best single method or any other ensemble method; (iii) for the best results, the Bayesian ensemble method should integrate all methods that satisfy a statistical test of normality on training data. The Bayesian ensemble model easily integrates all kinds of RNA-seq data and priors as well as new and existing inference methods.The paper categorizes and reviews state-of-the-art underlying methods, describes the Bayesian stacking ensemble approach in detail, and presents experimental results. The source code and data used will be available to the community. |
