Skip to article frontmatterSkip to article content

Appendix

Database [DVC,GDrive]. We will need a database to store our raw and geoprocessed data. In addition, we will need to store our model parameters and resulting figures.

Representation [Raster, Point Cloud]. We will need to represent our weather stations as point clouds and we may also need to represent them as Rasters. So a blog post will be decided to showcase how we can move between them.

Masks [Country, Land/Ocean]. We need to mask our data

Masked Likelihoods [GPD, TPP].

Sensitivity Analysis [MC, Gauss Approx, Taylor, Unscented, Moment Matching]

Numpyro + PPL I - Model [Prior, Likelihood, Posterior, Prior Predictive Posterior].

Numpyro + PPL II - Guide [MLE, MAP, Laplace, VI, MCMC, HMC].

Missing Data. e.g., Convolutions, Gaussian Processes, Masked-Likelihoods

CRS, Transform, Bounds, Resolution.


Preprocessing

Extreme Values [BM, POT, TPP]


Algorithms

Gaussian Processes [GP, Kriging, Kernel Methods].

Sparse Gaussian Processes [SGP].

Ensemble Kalman Filter [EnsKF]


Other



Algorithms


Architectures


ROM


Multiscale


Objective-Based Approaches


Conditional Generative Models


Engineering Tricks


Numpyro Tutorials


Keras Tutorial