Presentation
The exponential growth of data collection, particularly in medical contexts, raises new challenges for statistical analysis. Datavers is an interdisciplinary research team combining expertise in Public Health (CHU de Lille) and Statistics (Inria Lille) to address these challenges.
Our objective is to develop methodologies potentially integrable into a clinical decision support system grounded in heterogeneous health data (Electronic Health Records, national databases, IoT medical devices, and multi-omics data). The inherent complexity of medical data (tall datasets, high-dimensional variables, and space-time dependencies) drives us to develop new statistical learning methods alongside our clinical applications.
Our research is structured along three axes:
- Axis 1: Framework for clinical decision support based on heterogeneous health data
- Axis 2: Methods for space-time dependent data analysis
- Axis 3: Approaches for complex, high-dimensional and tall data