Skip to main content
. Author manuscript; available in PMC: 2023 Oct 13.
Published in final edited form as: Annu Rev Stat Appl. 2021 Nov 17;9(1):289–319. doi: 10.1146/annurev-statistics-040120-010930

Figure 9.

Figure 9

Example of group sparsity patterns of cMLP first layer weights with four first layer hidden units (H=4) and four input series (p=4) with maximum lag k=4. Differing sparsity patterns are shown for the three different structured penalties: (a) group lasso, (b) group sparse group lasso, and (c) hierarchical lasso. Abbreviation: cMLP, component-wise multilayer perceptron.