Full TitlePersonalized therapy for RhEumatic DIseases via machine learning
Human health and medicine are changing rapidly, pressing the development of machine learning techniques for automatic diagnosis and prognosis. Mining the large, feature- rich, heterogeneous, noisy, incomplete, and multivariate time series arising from electronic medical records (EMR) is a very hard problem; addressing it will require pushing the boundaries of machine learning research. This is exactly the aim of this project: to propose cutting-edge machine learning techniques to address the problems emerging from EMR mining. The envisaged methods will be ultimately applied in therapy selection of rheumatic diseases, where two big databases collected by the Portuguese Society of Rheumatology will be available. The study of this population with rheumatic diseases will help establishing public health strategies that will enhance treatment via tailored therapies while decreasing the overall costs and improving the quality of life of the patients.