Core: Bayesian outlier rejection
Filtering peaks is a trivial form of outlier rejection, but in practice tends to throw out too many peaks. We need to implement some form of Bayesian outlier rejection because the statistical reasons to throw out a peak or not are not always clear.