Impute Missing Variables

Augment a Random Forest with automatic imputation.

Author

Florian Pfisterer

Published

January 31, 2020

Prerequisites

This tutorial assumes familiarity with the basics of mlr3pipelines. Consult the mlr3book if some aspects are not fully understandable. It deals with the problem of missing data.

The random forest implementation in the package ranger unfortunately does not support missing values. Therefore, it is required to impute missing features before passing the data to the learner.

We show how to use mlr3pipelines to augment the ranger learner with automatic imputation.

We load the mlr3verse package which pulls in the most important packages for this example.

library(mlr3verse)
Loading required package: mlr3

We initialize the random number generator with a fixed seed for reproducibility, and decrease the verbosity of the logger to keep the output clearly represented.

set.seed(7832)
lgr::get_logger("mlr3")$set_threshold("warn")

Construct the Base Objects

First, we take an example task with missing values (diabetes) and create the ranger learner:

library(mlr3learners)

task = tsk("diabetes")
print(task)

── <TaskClassif> (768x9): Synthetic Diabetes ───────────────────────────────────────────────────────────────────────────
• Target: diabetes
• Properties: twoclass
• Features (8):
  • dbl (8): age, glucose, insulin, mass, pedigree, pregnant, pressure, triceps
• Target classes: pos (positive class, 36%), neg (64%)
learner = lrn("classif.ranger")
print(learner)

── <LearnerClassifRanger> (classif.ranger): Random Forest ──────────────────────────────────────────────────────────────
• Model: -
• Parameters: num.threads=1
• Packages: mlr3, mlr3learners, and ranger
• Predict Types: [response] and prob
• Feature Types: logical, integer, numeric, character, factor, and ordered
• Encapsulation: none (fallback: -)
• Properties: hotstart_backward, importance, missings, multiclass, oob_error, selected_features, twoclass, and weights
• Other settings: use_weights = 'use', predict_raw = 'FALSE'

We can now inspect the task for missing values. task$missings() returns the count of missing values for each variable.

task$missings()
diabetes      age  glucose  insulin     mass pedigree pregnant pressure  triceps 
       0        0        5      405       13        0        0       35      251 

Additionally, we can see that the ranger learner can not handle missing values:

learner$properties
[1] "hotstart_backward" "importance"        "missings"          "multiclass"        "oob_error"        
[6] "selected_features" "twoclass"          "weights"          

For comparison, other learners, e.g. the rpart learner can handle missing values internally.

lrn("classif.rpart")$properties
[1] "importance"        "missings"          "multiclass"        "selected_features" "twoclass"         
[6] "weights"          

Before we dive deeper, we quickly try to visualize two of the columns with missing values:

autoplot(task$clone()$select(c("insulin", "triceps")), type = "pairs")

Operators overview

An overview over implemented PipeOps for imputation can be obtained like so:

mlr_pipeops$keys("^impute")
[1] "imputeconstant" "imputehist"     "imputelearner"  "imputemean"     "imputemedian"   "imputemode"    
[7] "imputeoor"      "imputesample"  

Construct Operators

mlr3pipelines contains several imputation methods. We focus on rather simple ones, and show how to impute missing values for factor features and numeric features respectively.

Since our task only has numeric features, we do not need to deal with imputing factor levels, and can instead concentrate on imputing numeric values:

We do this in a two-step process: * We create new indicator columns, that tells us whether the value of a feature is “missing” or “present”. We achieve this using the missind PipeOp.

  • Afterwards, we impute every missing value by sampling from the histogram of the respective column. We achieve this using the imputehist PipeOp.

We also have to make sure to apply the pipe operators in the correct order!

imp_missind = po("missind")
imp_num = po("imputehist", affect_columns = selector_type("numeric"))

In order to better understand we can look at the results of every PipeOp separately.

We can manually trigger the PipeOp to test the operator on our task:

task_ext = imp_missind$train(list(task))[[1]]
task_ext$data()
     diabetes missing_glucose missing_insulin missing_mass missing_pressure missing_triceps
       <fctr>          <fctr>          <fctr>       <fctr>           <fctr>          <fctr>
  1:      neg         present         missing      present          present         present
  2:      pos         present         missing      present          present         missing
  3:      neg         present         missing      present          present         present
  4:      neg         present         missing      present          present         missing
  5:      neg         present         missing      present          present         present
 ---                                                                                       
764:      pos         present         missing      present          present         present
765:      pos         present         present      present          present         present
766:      neg         present         present      present          present         present
767:      neg         present         missing      present          present         missing
768:      neg         present         missing      present          present         missing

For imputehist, we can do the same:

task_ext = imp_num$train(list(task))[[1]]
task_ext$data()
     diabetes   age pedigree pregnant glucose   insulin  mass pressure  triceps
       <fctr> <num>    <num>    <num>   <num>     <num> <num>    <num>    <num>
  1:      neg    28    0.240        0     194 181.02651  28.5       50 45.00000
  2:      pos    44    0.828        8     129 129.30320  27.6       84 23.56236
  3:      neg    26    0.696        2     132 253.29451  38.2       72 42.00000
  4:      neg    39    0.192        6     184 599.13042  24.7       62 40.82062
  5:      neg    22    0.178        4     108 162.01169  31.6       58 32.00000
 ---                                                                           
764:      pos    43    0.346       10      92 233.45051  33.1       50 28.00000
765:      pos    26    1.162        0     181 474.00000  35.8       58 27.00000
766:      neg    60    0.460        6     130 540.00000  34.2      106 27.00000
767:      neg    24    0.206        2     146   8.83319  27.7       72 28.42715
768:      neg    41    0.187        6     115 138.42919  27.6       72 28.06583

This time we obtain the imputed data set without missing values.

task_ext$missings()
diabetes      age pedigree pregnant  glucose  insulin     mass pressure  triceps 
       0        0        0        0        0        0        0        0        0 

Putting everything together

Now we have to put all PipeOps together in order to form a graph that handles imputation automatically.

We do this by creating a Graph that copies the data twice, processes each copy using the respective imputation method and afterwards unions the features. For this we need the following two PipeOps : * copy: Creates copies of the data. * featureunion Merges the two tasks together.

graph = po("copy", 2) %>>%
  gunion(list(imp_missind, imp_num)) %>>%
  po("featureunion")

as a last step we append the learner we planned on using:

graph = graph %>>% po(learner)

We can now visualize the resulting graph:

graph$plot()

Resampling

Correct imputation is especially important when applying imputation to held-out data during the predict step. If applied incorrectly, imputation could leak info from the test set, which potentially skews our performance estimates. mlr3pipelines takes this complexity away from the user and handles correct imputation internally.

By wrapping this graph into a GraphLearner, we can now train resample the full graph, here with a 3-fold cross validation:

graph_learner = as_learner(graph)
rr = resample(task, graph_learner, rsmp("cv", folds = 3))
rr$aggregate()
classif.ce 
 0.2669271 

Missing values during prediction

In some cases, we have missing values only in the data we want to predict on. In order to showcase this, we create a copy of the task with several more missing columns.

dt = task$data()
dt[1:10, "age"] = NA
dt[30:70, "pedigree"] = NA
task_2 = as_task_classif(dt, id = "diabetes2", target = "diabetes")

And now we learn on task, while trying to predict on task_2.

graph_learner$train(task)
graph_learner$predict(task_2)

── <PredictionClassif> for 768 observations: ───────────────────────────────────────────────────────────────────────────
 row_ids truth response
       1   neg      neg
       2   pos      neg
       3   neg      neg
     ---   ---      ---
     766   neg      neg
     767   neg      neg
     768   neg      neg

Missing factor features

For factor features, the process works analogously. Instead of using imputehist, we can for example use imputeoor. This will simply replace every NA in each factor variable with a new value missing.

A full graph might the look like this:

imp_missind = po("missind", affect_columns = NULL, which = "all")
imp_fct = po("imputeoor", affect_columns = selector_type("factor"))
graph = po("copy", 2) %>>%
  gunion(list(imp_missind, imp_num %>>% imp_fct)) %>>%
  po("featureunion")

Note that we specify the parameter affect_columns = NULL when initializing missind, because we also want indicator columns for our factor features. By default, affect_columns would be set to selector_invert(selector_type(c("factor", "ordered", "character"))). We also set the parameter which to "all" to add indicator columns for all features, regardless whether values were missing during training or not.

In order to test out our new graph, we again create a situation where our task has missing factor levels. As the (diabetes) task does not have any factor levels, we use the famous (boston_housing) task.

# task_bh_1 is the training data without missings
task_bh_1 = tsk("boston_housing")

# task_bh_2 is the prediction data with missings
dt = task_bh_1$data()
dt[1:10, chas := NA][20:30, rm := NA]
task_bh_2 = as_task_regr(dt, id = "bh", target = "cmedv")

Now we train on task_bh_1 and predict on task_bh_2:

graph_learner = as_learner(graph %>>% po(lrn("regr.ranger")))
graph_learner$train(task_bh_1)
graph_learner$predict(task_bh_2)

── <PredictionRegr> for 506 observations: ──────────────────────────────────────────────────────────────────────────────
 row_ids truth response
       1  24.0 25.08769
       2  21.6 22.31111
       3  34.7 32.08418
     ---   ---      ---
     504  23.9 24.08323
     505  22.0 22.58690
     506  19.0 20.43212

Success! We learned how to deal with missing values in less than 10 minutes.