Madlene Nussbaum: Mastering ML for spatial prediction I - overview and introduction in methods

Many algorithm driven statistical methods are nowadays used for (spatial) prediction. Participants will get an overview of the different types/families of methods (shrinkage, generalized additive models, tree based methods, neural networks, support vector machines) and different machine learning concepts (bootstrap, boosting, model averaging). Three methods selected from different families (random forest, support vector machines, lasso) are presented in more detail including tuning of model parameters for these models.

Installation instructions
You need R and an Editor, e.g. RStudio
Following packages, run code:
"gbm", "geoGAM", "raster", "quantregForest"))

Documents

[dreznik]

this was more of an overview of basic ML algorithms than a talk about geospatial predictive modeling (only one slide on it!). suggestions: use h2o.automl() so algorithms are selected under the hood. also: do a 2nd talk on specifics of geospatial modeling, please!