-
|
Hello, using plot_decision_region, how can one view what the decision region looks like for examples that are outside of the trained distribution. For the XOR function book example, y(i) = 0 if x0(i) * x1(i) < 0 otherwise 1, with training values between (-1, 1), I'd like to view the decision region for test values 10x smaller and larger than (-1, 1). In the provided screenshot of a decision boundary created by 2000 (-1, 1) examples, a model.predict([[-2, .03]]) would result in a guess of 1 rather than 0. The decision boundary extends outward to larger numbers because even tests like (-100, 3) results in 1 rather than 0.
|
Beta Was this translation helpful? Give feedback.
Replies: 2 comments 1 reply
-
|
Good question, I think the |
Beta Was this translation helpful? Give feedback.
-
|
I was able to use the zoom_factor feature in the latest build of mlxtend. I was using the default mlxtend on Colab and just upgraded to the latest build and that worked! |
Beta Was this translation helpful? Give feedback.
-
|
Awesome, glad to hear! |
Beta Was this translation helpful? Give feedback.
Good question, I think the
zoom_factorparam could help you with this.