Representation is the process of mapping raw data into features or structures that a model can use effectively. Good representations make relevant patterns easier for a learning algorithm to discover and use.
The process of mapping data to useful features.