Displays the results of cross-validation to tune the number of components to
retain in each dataset for a sPLS run. Similar to
mixOmics::plot.tune.spls().
Arguments
- spls_tune_res
The result of
spls_tune()ormixOmics::tune.spls().
Details
The plot displays the correlation or RSS between the latent components
obtained with the corresponding values of keepX (x-axis) and keepY
(y-axis) and the latent components of the full model (i.e. that retains all
features). The colour of the points shows the mean correlation/RSS across the
cross-validation folds, and the size of the points' shadow (in gray)
represents the coefficient of variation (COV) of the correlation/RSS, i.e.
standard error divided by mean. The point corresponding with the optimal
value for keepX and keepY is indicating with a red border.