remember to manage the length of the predictions in `inner_loop()`/`spatial_prediction()`/`model_validation()`

idem-lab / map-ir-pipeline

Prototype/pseudo code demonstrating stacked generalisation method used in https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.3000633#sec010

0 stars 0 forks source link

remember to manage the length of the predictions in `inner_loop()`/`spatial_prediction()`/`model_validation()` #65

Open njtierney opened 1 month ago

njtierney commented 1 month ago

so we only get out length N out of sample predictions (the phenotypic predictions).

It might be easiest to pad out the genotypic (M) predictions with NA values.

With reference to this code:

  outer_loop_results_spatial = spatial_prediction(
    covariate_rasters = raster_covariates,
    training_data = ir_data_subset,
    level_zero_models = model_list,
    inla_mesh_setup = gp_inla_setup
  )