add note on development

vignettes/cast01-CAST-intro-cookfarm.Rmd

@@ -21,6 +21,8 @@ user_hanna <- Sys.getenv("USER") %in% c("hanna")
 
 
 ## Introduction
+!!Note: Some recent developments of CAST are not yet fully documented in this tutorial. A major update can be expected for Apr 2024!!
+
 ### Background
 One key task in environmental science is obtaining information of environmental variables continuously in space or in space and time, usually based on remote sensing and limited field data. In that respect, machine learning algorithms have been proven to be an important tool to learn patterns in nonlinear and complex systems.
 However, standard machine learning applications are not suitable for spatio-temporal data, as they usually ignore the spatio-temporal dependencies in the data. This becomes problematic in (at least) two aspects of predictive modelling: Overfitted models as well as overly optimistic error assessment (see [Meyer et al 2018](https://www.sciencedirect.com/science/article/pii/S1364815217310976) or [Meyer et al 2019](https://www.sciencedirect.com/science/article/abs/pii/S0304380019303230) ). To approach these problems, CAST supports the well-known caret package ([Kuhn 2018](https://topepo.github.io/caret/index.html) to provide methods designed for spatio-temporal data.