You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository was archived by the owner on Aug 1, 2024. It is now read-only.
First experiments show that neighbor based models will have a hard time coming up with a decent model for 6 features. It is also hard to get an intuition for how decisions are made.
It seems especially hard to fight overfitting. Just for the general intuition of overfitting, this can be useful
https://scikit-learn.org/stable/modules/neighbors.html#nearest-neighbors-classification
First experiments show that neighbor based models will have a hard time coming up with a decent model for 6 features. It is also hard to get an intuition for how decisions are made.
It seems especially hard to fight overfitting. Just for the general intuition of overfitting, this can be useful
https://scikit-learn.org/stable/auto_examples/model_selection/plot_underfitting_overfitting.html
For a first experiment we restrict ourselves to 2 features and use code like this
https://scikit-learn.org/stable/auto_examples/neighbors/plot_classification.html#sphx-glr-auto-examples-neighbors-plot-classification-py
to display the decision boundaries.
To come up with a good number of neighbors (and possibly other parameters) we could use hyper parameter search like the ones provided by https://scikit-learn.org/stable/modules/grid_search.html