Synthetic Data Increases Equity for Underrepresented Populations in Cancer Clinical Trials.

The article explains how synthetic minority over sampling techniques (SMOTE), a type of machine learning, can improve representation and equity for underrepresented populations in cancer clinical trials. It discusses the usefulness of SMOTE when creating or testing predictive models, as well as for balancing the distribution of minority class samples. Limitations of SMOTE include lack of accuracy of representation of original data and high computational costs with large data sets.