From: Evaluation of tree-based statistical learning methods for constructing genetic risk scores
Algorithm | Hyperparameter | Considered realizations |
---|---|---|
Random forests & random forests VIM | mtry | \(\left\lfloor \begin{pmatrix} 0.5&1&2 \end{pmatrix} \cdot \lfloor \sqrt{p} \rfloor \right\rfloor\) |
min.node.size | \(\left\lfloor \begin{pmatrix} 0.01&0.05&0.1 \end{pmatrix} \cdot N \right\rfloor\) | |
num.trees | 2000 | |
Logic regression & logic bagging | ntrees | \(\begin{pmatrix} 1&2&3&4&5&6 \end{pmatrix}\) |
nleaves | \(\begin{pmatrix} 1&2&\ldots&9&10 \end{pmatrix}\) (Simulation studies) | |
\(\begin{pmatrix} 1&2&\ldots&19&20 \end{pmatrix}\) (Real data application) | ||
Logic regression | Cooling schedule | Experimental |
Simulated annealing iterations | 500000 | |
Logic bagging | Bagging iterations | 500 |
Elastic net | \(\alpha\) | \(\begin{pmatrix} 0.5&0.75&0.9&0.99 \end{pmatrix}\) |
\(\lambda\) | Cross-validation |