Definition
The process of choosing the most relevant subset of input features from the original dataset to use in model training, reducing complexity and training time.
In short: The process of choosing the most relevant subset of input features from the original dataset to use in model training, reducing complexity and training time.