FIGURE 3.

Schematical overview of estimating the hyperparameters. Step 1: the global prior variance ${\tau }_{global}^2$ is estimated. Step 2: group weights ${\mathit{\gamma }}^{(d)}$ and hyperpenalties ${\lambda }^{(d)}$, $d=1,\dots ,D,$ are estimated for each co‐data set separately using appropriate shrinkage. Step 3: group set weights $\mathit{w}$ are estimated to combine the co‐data sets. The estimated hyperparameters are used to estimate the regression coefficients $\widehat{\mathit{\beta }}$ as given in Equation 3 [Colour figure can be viewed at wileyonlinelibrary.com]