Title: Distributed inference for generalized linear models with variable selection
Authors: Luca Maestrini - The Australian National University (Australia) [presenting]
Matias Quiroz - University of Technology Sydney (Australia)
Feng Li - Central University of Finance and Economics (China)
Abstract: A computationally efficient framework is proposed for fitting generalized linear models and performing variable selection through a distributed system. The dataset is partitioned into numerous subsets and maximum likelihood estimation is performed on each subset. The resulting estimates are combined together to form a pseudo-likelihood approximation to the likelihood function which can be linked to a penalization. The final parameter estimation is conducted via a fast approximate inference method. We illustrate our results on prominent generalized linear models using different priors for variable selection.