As far as validation measures are concerned, there are also several general studies, e.g., [ 18], aimed at establishing the intrinsic, as well as the relative, merit of a measure.
Note, that the clinical relevant quantification of the myocardial blood flow (MBF) could also be used as an additional validation measure.
For this reason and based also on an open problem mentioned in [ 20], we perform several experiments with different parameter settings of H and p, in order to find the "best" precision-time trade-off, when Consensus is regarded as an internal validation measure.
Finally, only a subset of our cohort had a diagnostic interview as the criteria validation measure, so further studies with larger sample sizes are required to confirm these findings.
As already outlined before, accuracy is not a suitable validation measure for highly unbalanced virtual screening datasets.
For the presented validation study, we selected the area under precision recall curve (AUPRC) as validation measure.
Therefore, and since this paper focuses on the presentation of the data and the motion compensation methods are included mostly to illustrate the utility of the data, we refrained from adding MBF as validation measure.
Our work shows that, as far as clustering and validation measures go, choices more convenient than NMF are possible: it is somewhat reductive to consider NMF as a clustering algorithm.
In the remainder of this paper, we concentrate on Consensus and FC as internal validation measures.
In this study, we have taken an initial step in evaluating convergent validity of a relative PA question using indicators of physical functioning as validation measures.
