Robust regression for time series exhibiting heteroscedasticity

Final year master’s degree Internship

Research
R
Statistical modeling
EDA
Author

Aurélien Callens

Published

September 1, 2018

Summary

During this internship, I worked on a new statistical method to perform robust regression for time series exhibiting heteroscedasticity. We developed and tested this method on a dataset containing chlorophyll concentration in a small tributary of the Thames River (UK).

Productions

  • My Master 2 Thesis:
  • Developement in the rlmDataDriven package:
    • rlmDD_het.R : this function performs a robust regression which accounts for temporal correlations and heterogeneity.
    • whm.R : this function is the R implementation of the weighted M-estimation.
  • A published paper in a reviewed journal:
    • Callens, A., Wang, Y., Fu, L. et al. (2020). Robust Estimation Procedure for Autoregressive Models with Heterogeneity. Environmental Modeling & Assessment, (10.1007/s10666-020-09730-w)
Back to top