K-Nearest Neighbors

Non parametric model used for both classification and regression tasks.

Hyper parameter K.

No explicit training phase, makes predictions based on the proximity of data points.

There is an optimal K value, and model performs poorly when you have a lot of features.