RTFBoost: robust tree-based functional boosting
In practical functional regression problems, measurements of the response and the functional predictor may either or both contain atypical values, which can cause serious damage to the regression estimator if not taken into consideration. We provide two options to robustify tree-based functional boosting, inspired by M and MM-estimators respectively, and investigate their performance through numerical experiments. We have also developed an R package called “RTFBoost” that implements our proposal.
Xiaomeng Ju
Postdoctoral research fellow in Biostatistics
I am a postdoctoral research fellow in the Division of Biostatistics, at the New York University, Grossman School of Medicine, advised by Professor Matias Salibian-Barrera. She received her BSc in Statistics from Renmin University of China, and MA in Statistics from University of Michigan. Xiaomeng’s research is centred on computational statistics with a special focus on robust statistics and functional data. Her ongoing thesis work develops gradient boosting methods for regression problems with complex data.