Norwegian version of this page

Statistical models for high-dimensional and functional data

The research group aims at developing statistical methods for analyzing data with complex structure, high-dimensional and/or functional (intrinsically smooth)

Image may contain: Colorfulness, Text, Line, Plot.

Various examples of high-dimensional data and analyses

About the group

The rapid evolution of technological devices in biomedicine allows measuring increasingly complex and high-dimensional data (large omics databases, high-frequency time series, 3D images generated by medical scanners). Proper data understanding and modeling is key to reaching meaningful conclusions, targeted at improving and personalizing clinical treatment. Indeed, data are a powerful source of information that can be leveraged jointly to structural knowledge to enhance the model power and accuracy, particularly in the presence of noise, incompleteness, and inconsistent patterns.

Advanced statistical methods provide the basic tools to extract knowledge from data, to quantify the associated uncertainty, and to predict or make decisions accounting for such uncertainty. Accounting for complexity requires the development of novel statistical methods and computational approaches, fueling a fascinating and fast-growing area of research in statistics.

Estimating fundamental structures and complex dependencies poses challenges at the frontiers of research in statistics, such as:

  • data incompleteness and fragmentation: we handle missing and unstructured data via latent variable modeling and careful propagation of uncertainties
  • complex data generating processes: we carefully exploit the data intrinsic structure or regularity via combination of advanced mathematical tools (functional spaces, networks, tensors) to achieve better modeling
  • redundancy in information: we develop sparse statistical models capable of capturing and summarizing the relevant insight from the data
  • heterogeneous data sources: we develop knowledge-based data integration to jointly analyze and exploit the several data modalities that allow a comprehensive and holistic description of the patient condition

Projects

Methods for high-dimensional and functional data have been flourishing over the last decades, and our aim is to contribute to the development in this field. We work on:

  • scalable (Bayesian) clustering and signature discovery for molecular data
  • multi-modal data integration for personalized treatment recommendations
  • rank-based statistical models for preference learning
  • sparse models and structural equation modeling for functional data
  • high-dimensional mixture cure models for survival data

Cooperation

The group cooperates with both national and international researchers:

Published Oct. 28, 2020 11:21 AM - Last modified Jan. 31, 2024 10:51 AM

Contact

Dept. of Biostatistics
Domus Medica, Gaustad
Sognsvannsveien 9
0372 Oslo

Group Leader

Valeria Vitelli

Participants

Detailed list of participants