There is an important clarification to be made here: the reviewers are thinking in terms of coordinate systems, with the difference between our and prior experiments simply being the point of origin, but this is not the case.
There is only one minimally sufficient adjustment set in the prior DAG, simply { Age, Comorbidity index}.
Will England pick Prior again, simply because Bairstow was so underwhelming in those last two Tests?
We will refer to the network using the prior policy simply as the Prior, and the network whose policy has since been modified as the Agent.
However, all is not lost if this didn't happen; the lack of a prior plan simply means that one needs to be established immediately.
Adding such a prior assumption simply states that we expect the state of the system to remain the same between successive time points; this is both parsimonious and would seem to be justified in the case of gene expression datasets, where we might only expect to observe a small number of transitions to differing states across the time series.
Christley et al. (2009) were also able to work with an inaccurate prior but they used an extra parameter (set by cross-validation) to weight the prior information so an inaccurate prior would simply be given less weight than an accurate one.
Here again, PriorPC performs well above PC-lite, even though priors were simply sampled between [0,0.5) or (0.5,1]. For each dataset, the more prior is included, the better the network can be recovered. This indicates that PriorPC is consistent.
In its prior incarnation, Sell Simply just didn't quite have the foundation and infrastructure to do that, but with help from his new Portland friends, things are looking up.
On the other hand, if the order is unknown and the cost function has to be calculated for a wide range of orders, e.g., L k = 1, 2,..., only a single update is required as the information from prior iterations are simply reused.
