Fitting Multidimensional Gaussian Process Regressions

Contents include:

• Simulating multidimensional data across location, age, and time
• Fitting GPs in GPFlow and TMB
• Testing scalability for larger dimensional models

Underlying Model to estimate:

Y_{c,a,t} = X^\beta + \Sigma  
\Sigma ~ GP({c} \kron {a} \kron {t})

Current implementations:

• GPFlow and ODVGP
• TMB

References:

• GPFlow
• ODVGP
• TMB