"constraint score" is a correct and usable expression in written English.
It is commonly used to refer to a numerical score given to a person or an entity based on how well they adhere to predetermined constraining parameters. For example, the students in the class were given a constraint score based on how closely they followed the teacher's instructions.
Features are selected through minimizing this constraint score, with maximizes class separability.
Constraint score is a supervised feature selection algorithm [9] which requires a relatively small amount of labeled data.
The constraint score approach selects features based on a small amount of labeled data but ignores unlabeled data.
We compare CCLS with unsupervised Laplacian score, as well as supervised constraint score, constrained Laplacian score, Spec, and ReliefF.
These include spectral feature selection (Spec) [27], ReliefF [28], Laplacian score, constraint score, and constrained Laplacian score.
It can be seen that the performance of CCLS is significantly better than that of Spec, Laplacian score, constraint score, and constrained Laplacian score.
In order to calculate constraint scores for Agrp exons, the STAT sites, and the IREs, we modified the concept of an RS score, such that the score for constraint represents the sum of constrained alignment columns in these regions while ignoring unconstrained alignment columns.
Better contact constraint scoring functions, which more evenly weight contact constraints according to both sequence separation and neighbourhood density, are needed to deal with this problem.
That there was a strong association of OHRQoL with negative emotionality and constraint scores supports the findings of Kressin et al [ 12], who observed a similar association with poor OHRQoL among older male US veterans.
This is probably a consequence of linkage to conserved sites and is likely to be associated with the way that GERP is calculated, since we find that sites with the most negative constraint scores tend to be preferentially located adjacent to runs of conserved sites in non-coding regions (for more details see Additional file 1: Figure S20).
Although these only represent a small fraction of the total number of pairs of individuals compared across all populations, it shows that the framework of comparing distributions of constraint scores on the individual level is able to detect significant differences, even for supposedly 'healthy' individuals from the 1000 Genomes populations.
