12:4012:40, 21 July 2019diffhist−84
K-nearest neighbors algorithm
Irrelevant to the model. Simplicity is by no way a considerable feature while choosing a statistical model. Generalization, robustness, computational complexity etc are relevant. BTW K-NN is expensive and complex O(n^2m). "All models are wrong but some are useful"
20:0820:08, 8 January 2016diffhist+16
Andrés W. López
Set forth term of Lopez' chairmanship of Harvard Law School Latino Alumni Committee. I am his predecessor and participated in appointing his successor.
12:4012:40, 21 July 2019diffhist−84
K-nearest neighbors algorithm
Irrelevant to the model. Simplicity is by no way a considerable feature while choosing a statistical model. Generalization, robustness, computational complexity etc are relevant. BTW K-NN is expensive and complex O(n^2m). "All models are wrong but some are useful"
20:0820:08, 8 January 2016diffhist+16
Andrés W. López
Set forth term of Lopez' chairmanship of Harvard Law School Latino Alumni Committee. I am his predecessor and participated in appointing his successor.