In order to avoid the curse of dimensionality when building a kNN model, which of the following options would be appropriate?
1) Feature selection
2) Feature extraction
3) Dimensionality reduction
4) All of the above