Using decision boundary display to plot the relationship between any 2 features if model is fitted to more than 2 features

[ETTAN93]

Describe the workflow you want to enable

Currently, it seems like it is not possible to pass in a model that has been fitted to more than 2 features to the DecisionBoundaryDisplay.from_estimator method. Is it possible to allow that while only passing in the 2 features you are interested in looking at the relationship for?

Describe your proposed solution

Allowing users to use decision boundary display to plot the relationship between any 2 features in a model that is fitted for more than 2 features.

Describe alternatives you've considered, if relevant

No response

Additional context

No response

[glemaitre]

Is the 2D partial dependence not better suited for your analysis?

[SuperG8571]

For example, when I have trained a support vector machine on the Iris dataset utilizing all 4 features, I want to display the decision boundary in a figure. However, the DecisionBoundaryDisplay.from_estimator method accepts 2 features at most. So if there is a method to project high dimensional models onto lower dimensional spaces instead of training new models. In my opinion, we may select the first 2 features from the 4 for calculation.
Additionally, it is puzzled to segment low dimensional points based on high dimensional surfaces. Because we need select a cross section to form a curve, then segment the points based on the curve. It is worth noting that the projection of a surface is a curve on the X, Y, Z plane is incorrect.
Therefore, is there any good way to display a model with more than 2 features?