Outline:

Fitting any model to a large-p-large-n dataset is always challenging computationally, if not statistically. The main aim of this project is to explore the techniques of random projection and sample splitting for fitting models to huge data when dimensionality and sample sizes are both beyond the capability of a single computer.