Hierarchical Canonical Correlation Analysis Reveals Phenotype, Genotype, and Geoclimate Associations in Plants
Visualizing predictions in comparison to the range of true values. The plots show how well the predicted values fit the ground-truth values in two cases: (a) using only gene expression data and (b) using all the four datasets by HCCA. In the plots, the -axis represents the samples sorted by the true FT10 values, and -axis the FT10 values. The ground-truth values are shown as the solid line, dashed lines represent a margin of 15%, and the dots represent the prediction of each sample. At the top, we also show the mean squared error (MSE) of the predictions.