Examples of additional tidymodels features

In this vignette, we illustrate how a number of features from tidymodels can be used to enhance a conventional SDM pipeline. We recommend users first become familiar with tidymodels; there are a number of excellent tutorials (both introductory and advanced) on its dedicated website We reuse the example on the Iberian lizard that we used in the tidysdm overview article.

Exploring models with DALEX

An issue with machine learning algorithms is that it is not easy to understand the role of different variables in giving the final prediction. A number of packages have been created to explore and explain the behaviour of ML algorithms, such as those used in tidysdm. In the tidysdm overview article, we illustrated how to use recipes to create profiles.

Here we demonstrate how to use DALEX, an excellent package that has methods to deal with tidymodels. tidysdm contains additional functions that allow use to use the DALEX functions directly on tidysdm ensembles.

We will use a simple ensemble that we built in the overview vignette.

library(tidysdm)
lacerta_ensemble
#> A simple_ensemble of models
#> 
#> Members:
#> • default_glm
#> • default_rf
#> • default_gbm
#> • default_maxent
#> 
#> Available metrics:
#> • boyce_cont
#> • roc_auc
#> • tss_max
#> 
#> Metric used to tune workflows:
#> • boyce_cont

The first step in DALEX is to create an explainer object, which can then be queried by different functions in the package, to turn the explainer into an explanation (following the DALEX lingo). As a first step, we use the custom function explain_tidysdm to generate our explainer:

explainer_lacerta_ens <- explain_tidysdm(lacerta_ensemble)
#> Preparation of a new explainer is initiated
#>   -> model label       :  data.frame  (  default  )
#>   -> data              :  444  rows  4  cols 
#>   -> data              :  tibble converted into a data.frame 
#>   -> target variable   :  444  values 
#>   -> predict function  :  predict_function 
#>   -> predicted values  :  No value for predict function target column. (  default  )
#>   -> model_info        :  package tidysdm , ver. 0.9.6.9004 , task classification (  default  ) 
#>   -> model_info        :  type set to  classification 
#>   -> predicted values  :  numerical, min =  0.0774875 , mean =  0.3221553 , max =  0.7481027  
#>   -> residual function :  difference between y and yhat (  default  )
#>   -> residuals         :  numerical, min =  -0.6701233 , mean =  -0.07215532 , max =  0.7139126  
#>   A new explainer has been created!

Now that we have our explainer, we can explore variable importance for the ensemble:

library(DALEX)
#> Welcome to DALEX (version: 2.4.3).
#> Find examples and detailed introduction at: http://ema.drwhy.ai/
#> Additional features will be available after installation of: ggpubr.
#> Use 'install_dependencies()' to get all suggested dependencies
#> 
#> Attaching package: 'DALEX'
#> The following object is masked from 'package:dplyr':
#> 
#>     explain
vip_ensemble <- model_parts(explainer = explainer_lacerta_ens)
plot(vip_ensemble)

Or generate partial dependency plots for a given variable (e.g. bio05):

pdp_bio05 <- model_profile(explainer_lacerta_ens, N = 500, variables = "bio05")
plot(pdp_bio05)

There are many other functions in DALEX that can be applied to the explainer to further explore the behaviour of the model; see several tutorial on https://modeloriented.github.io/DALEX/

It is also possible to explore the individual models that make up the ensemble:

explainer_list <- explain_tidysdm(tidysdm::lacerta_ensemble, by_workflow = TRUE)
#> Preparation of a new explainer is initiated
#>   -> model label       :  default_glm 
#>   -> data              :  444  rows  4  cols 
#>   -> data              :  tibble converted into a data.frame 
#>   -> target variable   :  444  values 
#>   -> predict function  :  yhat.workflow  will be used (  default  )
#>   -> predicted values  :  No value for predict function target column. (  default  )
#>   -> model_info        :  package tidymodels , ver. 1.2.0 , task classification (  default  ) 
#>   -> model_info        :  type set to  classification 
#>   -> predicted values  :  numerical, min =  0.2280177 , mean =  0.75 , max =  0.9854359  
#>   -> residual function :  difference between y and yhat (  default  )
#>   -> residuals         :  numerical, min =  -0.9096205 , mean =  5.395921e-12 , max =  0.7719823  
#>   A new explainer has been created!  
#> Preparation of a new explainer is initiated
#>   -> model label       :  default_rf 
#>   -> data              :  444  rows  4  cols 
#>   -> data              :  tibble converted into a data.frame 
#>   -> target variable   :  444  values 
#>   -> predict function  :  yhat.workflow  will be used (  default  )
#>   -> predicted values  :  No value for predict function target column. (  default  )
#>   -> model_info        :  package tidymodels , ver. 1.2.0 , task classification (  default  ) 
#>   -> model_info        :  type set to  classification 
#>   -> predicted values  :  numerical, min =  0.1315421 , mean =  0.7480648 , max =  1  
#>   -> residual function :  difference between y and yhat (  default  )
#>   -> residuals         :  numerical, min =  -0.6878921 , mean =  0.001935171 , max =  0.5870619  
#>   A new explainer has been created!  
#> Preparation of a new explainer is initiated
#>   -> model label       :  default_gbm 
#>   -> data              :  444  rows  4  cols 
#>   -> data              :  tibble converted into a data.frame 
#>   -> target variable   :  444  values 
#>   -> predict function  :  yhat.workflow  will be used (  default  )
#>   -> predicted values  :  No value for predict function target column. (  default  )
#>   -> model_info        :  package tidymodels , ver. 1.2.0 , task classification (  default  ) 
#>   -> model_info        :  type set to  classification 
#>   -> predicted values  :  numerical, min =  0.3390188 , mean =  0.7314788 , max =  0.9632964  
#>   -> residual function :  difference between y and yhat (  default  )
#>   -> residuals         :  numerical, min =  -0.9268645 , mean =  0.01852121 , max =  0.6280424  
#>   A new explainer has been created!  
#> Preparation of a new explainer is initiated
#>   -> model label       :  default_maxent 
#>   -> data              :  444  rows  4  cols 
#>   -> data              :  tibble converted into a data.frame 
#>   -> target variable   :  444  values 
#>   -> predict function  :  yhat.workflow  will be used (  default  )
#>   -> predicted values  :  No value for predict function target column. (  default  )
#>   -> model_info        :  package tidymodels , ver. 1.2.0 , task classification (  default  ) 
#>   -> model_info        :  type set to  classification 
#>   -> predicted values  :  numerical, min =  0.1095764 , mean =  0.6256817 , max =  0.9960248  
#>   -> residual function :  difference between y and yhat (  default  )
#>   -> residuals         :  numerical, min =  -0.8207859 , mean =  0.1243183 , max =  0.8904236  
#>   A new explainer has been created!

The resulting list can be then used to build lists of explanations, which can then be plotted.

profile_list <- lapply(explainer_list, model_profile,
  N = 500,
  variables = "bio05"
)
plot(profile_list)

The initial split

The standard approach in tidymodels is to make an initial split of the data into a test and a training set. We will use retain 20% of the data (1/5) for the testing set, and use the rest for training.

We start by loading a set of presences and absences and their associated climate, analogous to the one that we generated in the tidysdm overview article:

library(tidysdm)
library(sf)
lacerta_thin <- readRDS(system.file("extdata/lacerta_climate_sf.RDS",
  package = "tidysdm"
))

We then use spatial_initial_split to do the split, using a spatial_block_cv scheme to partition the data:

set.seed(1005)
lacerta_initial <- spatial_initial_split(lacerta_thin,
  prop = 1 / 5, spatial_block_cv
)
autoplot(lacerta_initial)

And check the balance of presences vs pseudoabsences:

check_splits_balance(lacerta_initial, class)
#> # A tibble: 1 × 4
#>   presence_test pseudoabs_test presence_train pseudoabs_train
#>           <int>          <int>          <int>           <int>
#> 1            88            267             25              72

We can now extract the training set from our lacerta_initial split, and sample folds to set up cross validation (note that we set the cellsize and offset based on the full dataset, lacerta_thin; this allows us to use the same grid we used for the initial_split).

set.seed(1005)
lacerta_training <- training(lacerta_initial)
lacerta_cv <- spatial_block_cv(lacerta_training,
  v = 5,
  cellsize = grid_cellsize(lacerta_thin),
  offset = grid_offset(lacerta_thin)
)
autoplot(lacerta_cv)

And check the balance in the dataset:

check_splits_balance(lacerta_cv, class)
#> # A tibble: 5 × 4
#>   presence_assessment pseudoabs_assessment presence_analysis pseudoabs_analysis
#>                 <int>                <int>             <int>              <int>
#> 1                  74                  197                14                 70
#> 2                  59                  225                29                 42
#> 3                  73                  220                15                 47
#> 4                  76                  209                12                 58
#> 5                  70                  218                18                 49

Different recipes for certain models

Only certain type of models (e.g. glm, svm) struggle with correlated variables; other algorithms, such as random forests, can handle correlated variables. So, we will create two recipes, one with all variables, and one only with the variables that are uncorrelated:

lacerta_rec_all <- recipe(lacerta_thin, formula = class ~ .)
lacerta_rec_uncor <- lacerta_rec_all %>%
  step_rm(all_of(c(
    "bio01", "bio02", "bio03", "bio04", "bio07", "bio08",
    "bio09", "bio10", "bio11", "bio12", "bio14", "bio16",
    "bio17", "bio18", "bio19", "altitude"
  )))

lacerta_rec_uncor
#> 
#> ── Recipe ──────────────────────────────────────────────────────────────────────
#> 
#> ── Inputs
#> Number of variables by role
#> outcome:    1
#> predictor: 20
#> coords:     2
#> 
#> ── Operations
#> • Variables removed: all_of(c("bio01", "bio02", "bio03", "bio04", "bio07",
#>   "bio08", "bio09", "bio10", "bio11", "bio12", "bio14", "bio16", "bio17",
#>   "bio18", "bio19", "altitude"))

And now use these two recipes in a workflowset (we will keep it small for computational time), selecting the appropriate recipe for each model. We will include a model (polynomial support vector machines, or SVM) which does not have a wrapper in tidysdm for creating a model specification. However, we can use a standard model spec from yardstick:

lacerta_models <-
  # create the workflow_set
  workflow_set(
    preproc = list(
      uncor = lacerta_rec_uncor, # recipe for the glm
      all = lacerta_rec_all, # recipe for the random forest
      all = lacerta_rec_uncor # recipe for svm
    ),
    models = list(
      # the standard glm specs
      glm = sdm_spec_glm(),
      # rf specs with tuning
      rf = sdm_spec_rf(),
      # svm specs with tuning
      svm = parsnip::svm_poly(
        cost = tune(),
        degree = tune()
      ) %>%
        parsnip::set_engine("kernlab") %>%
        parsnip::set_mode("classification")
    ),
    # make all combinations of preproc and models,
    cross = FALSE
  ) %>%
  # tweak controls to store information needed later to create the ensemble
  # note that we use the bayes version as we will use a Bayes search (see later)
  option_add(control = stacks::control_stack_bayes())

We can now use the block CV folds to tune and assess the models. Note that there are multiple tuning approaches, besides the standard grid method. Here we will use tune_bayes from the tune package (see its help page to see how a Gaussian Process model is used to choose parameter combinations).

This tuning method (as opposed to use a standard grid) does not allow for hyper-parameters with unknown limits, but mtry for random forest is undefined as its upper range depends on the number of variables in the dataset. So, before tuning, we need to finalise mtry by informing the set dials with the actual data:

rf_param <- lacerta_models %>%
  # extract the rf workflow
  extract_workflow("all_rf") %>%
  # extract its parameters dials (used to tune)
  extract_parameter_set_dials() %>%
  # give it the predictors to finalize mtry
  finalize(x = st_drop_geometry(lacerta_thin) %>% select(-class))

# now update the workflowset with the new parameter info
lacerta_models <- lacerta_models %>%
  option_add(param_info = rf_param, id = "all_rf")

And now we can tune the models:

set.seed(1234567)
lacerta_models <-
  lacerta_models %>%
  workflow_map("tune_bayes",
    resamples = lacerta_cv, initial = 8,
    metrics = sdm_metric_set(), verbose = TRUE
  )
#> i    No tuning parameters. `fit_resamples()` will be attempted
#> i 1 of 3 resampling: uncor_glm
#> ✔ 1 of 3 resampling: uncor_glm (274ms)
#> i 2 of 3 tuning:     all_rf
#> ! No improvement for 10 iterations; returning current results.
#> ✔ 2 of 3 tuning:     all_rf (16s)
#> i 3 of 3 tuning:     all_svm
#> ✔ 3 of 3 tuning:     all_svm (18.7s)

We can have a look at the performance of our models with:

autoplot(lacerta_models)

Stack ensembles

Instead of building a simple ensemble with the best version of each model type, we can build a stack ensemble, as implemented in the package stacks. Stacking uses a meta-learning algorithm to learn how to best combine multiple models, including multiple versions of the same algorithm with different hyper-parameters.

library(stacks)
set.seed(1005)
lacerta_stack <-
  # initialize the stack
  stacks() %>%
  # add candidate members
  add_candidates(lacerta_models) %>%
  # determine how to combine their predictions
  blend_predictions() %>%
  # fit the candidates with non-zero weights (i.e.non-zero stacking coefficients)
  fit_members()

autoplot(lacerta_stack, type = "weights")

We can see that three versions of the SVM and one of the random forests were selected; the stacking coefficients give an indication of the weight each model carries within the ensemble. We can now use the ensemble to make predictions about the testing data:

lacerta_testing <- testing(lacerta_initial)

lacerta_test_pred <-
  lacerta_testing %>%
  bind_cols(predict(lacerta_stack, ., type = "prob"))

And look at the goodness of fit using some commonly used sdm metrics. Note that sdm_metric_set is first invoked to generate a function (with empty ()) that is then used on the data.

sdm_metric_set()(data = lacerta_test_pred, truth = class, .pred_presence)
#> # A tibble: 3 × 3
#>   .metric    .estimator .estimate
#>   <chr>      <chr>          <dbl>
#> 1 boyce_cont binary         0.853
#> 2 roc_auc    binary         0.986
#> 3 tss_max    binary         0.92

We can now make predictions with this stacked ensemble. We start by extracting the climate for the variables of interest

download_dataset("WorldClim_2.1_10m")
climate_vars <- lacerta_rec_all$var_info %>%
  filter(role == "predictor") %>%
  pull(variable)

climate_present <- pastclim::region_slice(
  time_ce = 1985,
  bio_variables = climate_vars,
  data = "WorldClim_2.1_10m",
  crop = iberia_poly
)
prediction_present <- predict_raster(lacerta_stack, climate_present,
  type = "prob"
)
library(tidyterra)
ggplot() +
  geom_spatraster(data = prediction_present, aes(fill = .pred_presence)) +
  scale_fill_terrain_c() +
  # plot presences used in the model
  geom_sf(data = lacerta_thin %>% filter(class == "presence"))

Using multi-level factors as predictors

Most machine learning algorithms do not natively use multilevel factors as predictors. The solution is to create dummy variables, which are binary variables that represent the levels of the factor. In tidymodels, this is done using the step_dummy function.

Let’s create a factor variable with 3 levels based on altitude.

library(tidysdm)
# load the dataset
lacerta_thin <- readRDS(system.file("extdata/lacerta_climate_sf.RDS",
  package = "tidysdm"
))
# create a topography variable with 3 levels based on altitude
lacerta_thin$topography <- cut(lacerta_thin$altitude,
                breaks = c(-Inf, 200, 800, Inf),
                labels = c("plains", "hills", "mountains"))
table(lacerta_thin$topography)
#> 
#>    plains     hills mountains 
#>        82       233       137

We then create the recipe by adding a step to create dummy variables for the topography variable.

# subset to variable of interest
lacerta_thin <- lacerta_thin %>% select(class, bio05, bio06, bio12, bio15, topography)

lacerta_rec <- recipe(lacerta_thin, formula = class ~ .) %>%
  step_dummy(topography)
lacerta_rec
#> 
#> ── Recipe ──────────────────────────────────────────────────────────────────────
#> 
#> ── Inputs
#> Number of variables by role
#> outcome:   1
#> predictor: 5
#> coords:    2
#> 
#> ── Operations
#> • Dummy variables from: topography

Let’s us see what this does:

lacerta_prep <- prep(lacerta_rec)
summary(lacerta_prep)
#> # A tibble: 9 × 4
#>   variable             type      role      source  
#>   <chr>                <list>    <chr>     <chr>   
#> 1 bio05                <chr [2]> predictor original
#> 2 bio06                <chr [2]> predictor original
#> 3 bio12                <chr [2]> predictor original
#> 4 bio15                <chr [2]> predictor original
#> 5 X                    <chr [2]> coords    original
#> 6 Y                    <chr [2]> coords    original
#> 7 class                <chr [3]> outcome   original
#> 8 topography_hills     <chr [2]> predictor derived 
#> 9 topography_mountains <chr [2]> predictor derived

We have added two “derived” variables, topography_hills and topography_mountains, which are binary variables that allow us to code topography (with plains being used as the reference level, which is coded by both hills and mountains being 0 for a given location). We can look at the first few rows of the data to see the new variables by baking the recipe:

lacerta_bake <- bake(lacerta_prep, new_data = lacerta_thin)
glimpse(lacerta_bake)
#> Rows: 452
#> Columns: 9
#> $ bio05                <dbl> 30.50350, 25.28050, 23.67800, 29.68875, 26.34075,…
#> $ bio06                <dbl> 1.477000, 3.631750, 0.789500, 6.048750, 1.869000,…
#> $ bio12                <dbl> 596, 1490, 1395, 729, 1324, 1409, 1260, 1390, 116…
#> $ bio15                <dbl> 50.59533, 50.07437, 47.24211, 58.88199, 51.62960,…
#> $ X                    <dbl> -5.394226, -8.374844, -7.886102, -8.231414, -7.17…
#> $ Y                    <dbl> 39.48495, 41.97207, 41.89992, 39.49710, 41.78401,…
#> $ class                <fct> presence, presence, presence, presence, presence,…
#> $ topography_hills     <dbl> 1, 1, 0, 0, 1, 1, 1, 1, 1, 0, 0, 1, 1, 1, 1, 0, 0…
#> $ topography_mountains <dbl> 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1…

We can now run the sdm as usual:

# define the models
lacerta_models <-
  # create the workflow_set
  workflow_set(
    preproc = list(default = lacerta_rec),
    models = list(
      # the standard glm specs
      glm = sdm_spec_glm(),
      # rf specs with tuning
      rf = sdm_spec_rf()
    ),
    # make all combinations of preproc and models,
    cross = TRUE
  ) %>%
  # tweak controls to store information needed later to create the ensemble
  option_add(control = control_ensemble_grid())
# tune
set.seed(100)
lacerta_cv <- spatial_block_cv(lacerta_thin, v = 3)
lacerta_models <-
  lacerta_models %>%
  workflow_map("tune_grid",
    resamples = lacerta_cv, grid = 3,
    metrics = sdm_metric_set(), verbose = TRUE
  )
#> i    No tuning parameters. `fit_resamples()` will be attempted
#> i 1 of 2 resampling: default_glm
#> ✔ 1 of 2 resampling: default_glm (193ms)
#> i 2 of 2 tuning:     default_rf
#> i Creating pre-processing data to finalize unknown parameter: mtry
#> ✔ 2 of 2 tuning:     default_rf (981ms)
# fit the ensemble
lacerta_ensemble <- simple_ensemble() %>%
  add_member(lacerta_models, metric = "boyce_cont")

We can now verify that the dummy variables were used by extracting the model fit from one of the models in the ensemble:

lacerta_ensemble$workflow[[1]] %>% extract_fit_parsnip()
#> parsnip model object
#> 
#> 
#> Call:  stats::glm(formula = ..y ~ ., family = stats::binomial, data = data)
#> 
#> Coefficients:
#>          (Intercept)                 bio05                 bio06  
#>            -6.920024              0.635493             -0.334427  
#>                bio12                 bio15      topography_hills  
#>            -0.002505             -0.113106             -1.851098  
#> topography_mountains  
#>            -2.531259  
#> 
#> Degrees of Freedom: 451 Total (i.e. Null);  445 Residual
#> Null Deviance:       508.4 
#> Residual Deviance: 189.5     AIC: 203.5

We can see that we have coefficients for topography_hills and topography_mountains.

Let us now predict the presence of the lizard in the Iberian Peninsula using the ensemble. Note that, for predict_raster() to work, the name and levels for a categorical variable need to match with those used when training the models (i.e. in the recipe with step_dummy()):

climate_present <- terra::readRDS(system.file("extdata/lacerta_climate_present_10m.rds",
                                              package = "tidysdm"))
# first we add a topography variable to the climate data
climate_present$topography <- climate_present$altitude
climate_present$topography <- terra::classify(climate_present$topography, rcl = c(-Inf, 200, 800, Inf), include.lowest=TRUE, brackets=TRUE)
library(terra)
levels(climate_present$topography)  <- data.frame(ID = c(0,1,2), topography = c("plains", "hills", "mountains"))
# now we can predict
predict_factor <- predict_raster(lacerta_ensemble, climate_present)
plot(predict_factor)