Closed stineb closed 2 years ago
May adopt what @fgiardin did as described in Giardina et al. in prep.:
ET was gap-filled with single-layer neural networks, using temperature, PAR, VPD and ET simulated by the SPLASH model as predictors (Davis et al., 2017). To build this model, we used the R package ‘NNET’ (Venables & Ripley, 2002) and ‘CARET’ (Kuhn et al., 2021), and used a neural network with a single hidden layer, 20 nodes, 10-fold cross-validated
@fgiardin please coordinate with @khufkens to (possibly) adopt this
Won't fix, not in line with the PLUMBER workflow - currently relying on the default plumber code.
We need data with standardized aggregation to daily, monthly and annual values.