Group observations into a fixed number (k) of clusters such that cluster members are more similar to each other than to observations in other clusters.
Artificial neural networks are used to model complex relationships between inputs and outputs by learning a non-linear function without manual feature engineering.