In order to address the problem of the curse of dimensionality, this work proposes essentially variance-based adaptive strategies aiming to build a cheap meta-model (i.e. surrogate model) by employing the sparse PDD approach with its coefficients computed by regression.
Essentially, the variance components analysis for the analytes distinguishes the pervasive sources of experimental variation from variation that is specific to the test material.
Sampling in small blocks within days resulted in essentially unbiased variance components.
Although the same patterns did not hold for worms fed UV-killed E. coli, a major caveat is that they displayed essentially no variance in early reproduction (Additional file 1: Table S6); correlations involving this trait must therefore be interpreted with caution.
For quantitative traits, the framework is essentially Analysis of Variance (ANOVA) that decomposes the variability in the trait amongst the different factors.
The repeated 10-fold cross-validation essentially eliminates split variance by using many splits and averaging over them, and furthermore it reduces internal variance by using more sample for training than holdout (under typical training-testing split ratios).
This second analysis essentially removed the variance in infection outcomes solely caused by host differences and focused on the virus control on the residual variance.
If the mean is close to zero, the variance essentially determines the shape of the histogram.
