dm.cs.tu-dortmund.de/mlbits/class-knn-discussion/
Nearest Neighbors: Discussion – Lecture Notes
consistent (neighbors tend to have similar values), for \(N,k\rightarrow \infty\) and \(k/N\rightarrow 0\) , we always get excellent predictions. [ CoHa67 ; HaTiFr09 ]
But: we only have finite training data [...] \(y\in \{ +1,-1\}\) , predict
\[\begin{align*} \hat{y}(x) =& \begin{cases} +1 & \text{if } f(x) \geq 0 \\ -1 & \text{otherwise} \end{cases} \end{align*}\]
Note: we can use this form of regression with n …
