Stat4Reg is a research laboratory at Umeå University, where scientists work at developing statistical and machine learning models and methods, as well as free software, for the analysis of large individual databases, including linkages of health, socio-economics, demographic and other administrative registers (record linkage data), longitudinal studies (brain imaging, intervention programs, panel studies, ...), etc.
Stat4Reg at Github: github.com/stat4reg
Marie Eriksson and Xavier de Luna have been awarded funding by the Swedish Research Council for two projects starting 2025 on "Integrated machine and deep learning models for predictive analysis in complex disease domains" (PI Marie Eriksson, Co-applicant Xijia Liu) and "Generative AI to promote open science in register based research" (PI Xavier de Luna, Co-applicant Anna Baranowska-Rataj).
Maria Josefsson and Xavier de Luna are organizing invited sessions at CMStat23 in Berlin, December 2023. The former on "Biostatistical methods in Alzheimer's disease and aging research" and the latter on "Distributional shifts and applications to missing data and causal inference".
New open PhD student position in collaboration with CEDAR (Center for Demographic and Ageing Research) with a project on "Studies of heterogeneous effects of critical life course events using machine learning". More information and how to apply: link. Deadline for application 13 of November 2023.
Three projects have been awarded funding by the Swedish Research Council this Fall: "Dynamic relations between cardiovascular, social and behaviourial risk factors of cognitive ageing"; "Innovative machine learning methods for comparison of predictions and outcomes in register data"; "Robots with Causal Capabilities".
Maria Josefsson is invited speaker at the International Biometric Conference in Riga, 10-15 July, in the session "Recent developments in probabilistic machine learning methods for causal inference"; and Xavier de Luna is invited speaker at the meeting Challenges for Categorical Data Analysis in Perugia, 12-13 May.
Marie Eriksson is moderator for the session "Potentials for AI in National Quality Registers" at the "Quality Registers for research Day" in Stockholm, 5 May.
The project "Machine learning to study causality with big datasets: towards methods yielding valid statistical conclusions" has been awarded funding from the Wallenberg Foundation and from the Swedish Research Council.
We are organizing two invited sesssions "Causal inference with machine learning" and "Causal mediation analysis" at the 14th CMStatistics 2021 meeting, 18-20 December at King's College London.
We are organizing an invited session on "Causal inference and register data" at the 8th Nordic-Baltic Biometrics Virtual Conference (Helsinki), June 7-10, 2021.
Open reseach assistant position (5 months): announcement (in Swedish)
Several open positions in Statistics and Data Science: Postdocs (2-year) - PhD program (4-year)
Stat4Reg researchers are organizing three invited sessions at the Virtual CMStatistics, December 2020.
New PhD student position (causal mediation analysis)
Stat4Reg members organized two invited sessions and gave four talks at the CMStatistics meeting in London, December 2019; on topics related to machine learning for causal inference and missing data problems. See program.
Stat4Reg team participating at the brännboll "World Championship" for the third year in a row: picture!
Seminar: Gentle introduction to mediation analysis given to researchers working with register data at the Umeå SIMSAM Lab. Slides here.
Job: We are looking for a programmer (C++, Python, parallel programming, etc) to work with our researchers at implementing novel machine learning methods
PhD defence: Methods for longitudinal brain imaging studies with dropout - Feb 2019
Conference on Machine Learning — Registration deadline 15 January
Research funding from Swedish research councils
New funding from Riksbankens Jubileumsfond
PhD defence: Methods for improving covariate balance in observational studies
New funding to Stat4Reg from the Swedish Research Council for 2017-2022
