KNN

K-means is the unsupervised version of K-NN KNN regression is the mean of the KNN group.

Problems: The models stores all the data. Each time an instance is added, it makes a distance check. Exponential - curse of dimensionality

If you had a feature space in one variable and an algorithm covers it 99%, the more dimensions we add, the overall coverage decreases exponentially

For example 2 features : 99*99 = 98% 100 features features : 99^100 : 36%

Requirement : Distance metric has meaning. For ex: housing prices nearby are not exactly similar. So KNN does not work everywhere