The aim of this axis is to propose original machine and statistical learning methods, including deep learning approaches to account for low-sample-size high-dimension setting. In particular, we consider modeling complex patient care trajectories for prognosis prediction and decision making, as well as, approaches using synthetic patient generation.
We focus on the supervised classification and prediction tasks (from a computer science point of view) and their applications in (i) clinical practice by guiding diagnostics (i.e., the identification of patients with undiagnosed conditions), (ii) prognostics (i.e., the predicting the outcomes of patients having a particular condition), and (iii) in providing insights to biomedicine (e.g., identifying how patient should be treated through features selection).
Keywords: Data-driven medicine, Model-based medicine, Learning health system, Small samples high dimensional data, Electronic Health Records, Machine learning, Bayesian inference
Seminal references:
T. Lartigue, S. Bottani, S. Baron, O. Colliot, S. Durrleman and S. Allassonnière. Gaussian Graphical Model exploration and selection in high dimension low sample size setting. IEEE Transactions on Pattern Analysis and Machine Intelligence. DOI: 10.1109/TPAMI.2020.2980542
Boulet S, Ursino M, Thall P, Jannot A‐S, Zohar S. (2019) Bayesian variable selection based on clinical relevance weights in small sample studies—Application to colon cancer. Statistics in Medicine, 38: 2228– 2247. DOI: 10.1002/sim.8107
Charles-Nelson A, Katsahian S, Schramm C. (2019) How to analyze and interpret recurrent events data in the presence of a terminal event: An application on readmission after colorectal cancer surgery. Statistics in Medicine, Aug 15;38(18):3476-3502. DOI: 10.1002/sim.8168
Coulet A., Shah N.H., Wack M., Chawki M.B., Jay N., Dumontier M. (2018). Predicting the need for a reduced drug dose, at first prescription. Scientific reports, 8(1), 1-11. DOI: 10.1038/s41598-018-33980-0