[x] "The plot shows the surrogate model for the chosen iteration."
[x] "shows the graph of the surrogate model and the acquisition function (infill criterion) ..."
[x] Link to Christophs book is broken
[x] "Higher dimensional search space: Shows the Partial Dependence Plot (PDP) of the surrogate model with regards to the chosen search space component. At iteration i, the marginal effect of the chosen search space component on the predicted outcome \hat y^{(i)} of the surrogate model at iteration i is computed and displayed. "
[x] "Identify the influence of a search space component on the prediction of ŷ (i): A partial dependence plot can show whether the relationship between the target and a search space component is linear, monotonic or more complex."
[x] Das würde ich rauslassen: "Identify quality of surrogate at single iteration: If the PDP shows a constant at iteration i the search space component under review does not influence ŷ (i) thus the surrogate might not be suited for this problem." Das ist so nicht korrekt; "konstant" kann heißen, dass das surrogate nicht ordentlich fitted, ODER, dass der parameter tatsächlich nicht wirklich einen Einfluss hat.
Description Plot surrogate and infill criterion for a single iteration.
Open topics