The mFIMMG Dataset stores a 10-year (2010-2019) activity collected by 6 general practitioners, and consists of 14175 patients and 6 main fields. The demographic field is composed by age and gender. The monitoring field (i.e., diastolic and systolic blood pressure, height, weight, and waist) contains only continuous predictors, as well as the lab tests field where all the laboratory outcomes are stored. The remaining fields such as pathologies, drugs, exam prescriptions are all categorical.
The mFIMMG dataset has been utilised to develop the work entitled:
'A Semi-Supervised Multi-Task Learning Approach for Predicting Short-Term Kidney Desease Evolution'
published to IEEE Journal of Biomedical and Health Informatics by Michele Bernardini, Luca Romeo, Emanuele Frontoni and Massih-Reza Amini.
The code to replicate our experimental procedure is available upon request.
Please send an email to vrai@dii.univpm.it for requesting the mFIMMG dataset.