Data splitting is when data is divided into two or more [[subset]]s. Typically, with a two-part split, one part is used to evaluate or test the data and the other to [[train]] the [[model]]. Data splitting is an important aspect of [[data science]], particularly for creating models based on data.