The considered stroke dataset (DOI:10.17632/x8ygrw87jw.1, DOI:10.1016/j.artmed.2019.101723) was pre-processed by removing attributes with more than 30% missing values, by converting categorical descriptive variables into numerical ones through one-hot encoding. After these steps, the dataset is described through 1 id (the first column), 13 descriptive variables, and the target variable (the last column). The dataset contains instances representing patients who have not relapsed to stroke (the majority class, last_column=0), and instances representing patients who had a recurrent stroke (the minority class, last_column=1). The sepsi dataset (DOI:10.1038/s41598-020-73558-3) is described through 1 id (the first column), 3 descriptive variables, and the target variable (the last column). The task is to predict if the patient survived or not, due to sepsis. The dataset consists of patients died because of sepsis (last_column=1), while the remaining survived (last_column=0). The target variable of the two datasets was aligned so that the label that indicates the relapse to stroke in the cerebral stroke dataset corresponds to the label indicating people who did not survive in the sepsis dataset. Finally, a reduced version was built by: i) imposing a balanced class distribution (obtained by downsampling the majority class); ii) further reducing the source domain dataset through a 10% stratified random sampling in order to facilitate the competitors to run the experiments to perform a comparative analysis.