Their experimentation indicates that the AIS-based classification paradigm has the intrinsic property of dealing more efficiently with highly skewed datasets.
Skewed datasets are known to complicate both learning and performance evaluation [ 52].
In this study the "reasonable proxy" relationship is evaluated across a highly skewed daily precipitation dataset in Minnesota for H1,H2 and H3.
To create the simulation data, we first applied iASeq to the real GM12878 data to identify 86,353 SNPs that were not skewed in any dataset using Pr(a i =0| X i, N i, Π, V, W )>0.5 as cutoff.
On the technical side, it has to be borne in mind that our knowledge of protein interactions is certainly incomplete and may contain many false positive interactions [ 55, 56] and the employed technologies may skew the datasets towards particular interactions [ 57].
The distributions are positively and negatively skewed for both datasets.
The distribution of over 100,000 IPIs over the entire dataset is skewed to longer IPIs, as has been noted previously [ 10, 11].
Furthermore, previous studies have shown that the log (GI50) values in the NCI-60 dataset are skewed toward certain thresholds (Lee et al., 2008).
One can then infer whether each SNP i is skewed in each individual dataset d based on the posterior probability Pr (b id, c id | X i, N i, Π, V, W ) = ∑ a i Pr (a i, b id, c id | X i, N i, Π, V, W ) after summing over all possible values of a i (see Additional file 1: Supplemental Method S.4 equation S.14).
