control_variables

  • Available in: GLM

  • Hyperparameter: no

Description

Control variables are special predictors that are included during model training but automatically excluded during inference/scoring. This feature allows you to account for certain factors during training without having them affect predictions.

Common use cases include:

  • Accounting for batch effects or experimental conditions

  • Controlling for confounding variables

  • Incorporating fixed effects that won’t be available at prediction time

When control variables are specified, GLM will exclude them during scoring. Model metrics and scoring history are calculated for both the restricted model (with control variables excluded) and the unrestricted model (with control variables included).

To get the unrestricted model with its own metrics use glm.make_unrestriced_glm_model()/h2o.make_unrestricted_glm_model(glm).

The control variables’ coefficients are set to zero in the variable importance table. Use the unrestricted model to get the variable importance table with all variables included.

Notes:

  • This option is experimental.

  • This option is applicable only for regression and binomial distribution.

  • This option is not available when cross validation is enabled.

  • This option is not available when Lambda search is enabled.

  • This option is not available when interactions are enabled.

Example

R

Python

        library(h2o)
        h2o.init()
        # import the airlines dataset:
        # This dataset is used to classify whether a flight will be delayed 'YES' or not "NO"
        # original data can be found at http://www.transtats.bts.gov/
        airlines <-  h2o.importFile("http://s3.amazonaws.com/h2o-public-test-data/smalldata/airlines/allyears2k_headers.zip")

        # convert columns to factors
        airlines["Year"] <- as.factor(airlines["Year"])
        airlines["Month"] <- as.factor(airlines["Month"])
        airlines["DayOfWeek"] <- as.factor(airlines["DayOfWeek"])
        airlines["Cancelled"] <- as.factor(airlines["Cancelled"])
        airlines['FlightNum'] <- as.factor(airlines['FlightNum'])

        # set the predictor names and the response column name
        predictors <- c("Origin", "Dest", "Year", "UniqueCarrier", "DayOfWeek", "Month", "Distance", "FlightNum")
        response <- "IsDepDelayed"

        # split into train and validation
        airlines_splits <- h2o.splitFrame(data =  airlines, ratios = 0.8)
        train <- airlines_splits[[1]]
        valid <- airlines_splits[[2]]

        # try using the `control_variables` parameter:
        airlines_glm <- h2o.glm(family = 'binomial', x = predictors, y = response, training_frame = train,
                validation_frame = valid,
                remove_collinear_columns = FALSE,
                score_each_iteration = TRUE,
                generate_scoring_history = TRUE,
                control_variables = c("Year", "DayOfWeek"))

        # print the AUC for the validation data
        print(h2o.auc(airlines_glm, valid = TRUE))

        # take a look at the coefficients_table
        airlines_glm@model$coefficients_table

        # take a look at the learning curve
        h2o.learning_curve_plot(airlines_glm)

        # get the unrestricted GLM model
        unrestricted_airlines_glm <- h2o.make_unrestricted_glm_model(airlines_glm)

# get variable importance
varimp <- h2o.varimp(airlines_glm)
varimp_unrestricted <- h2o.varimp(unrestricted_airlines_glm)

        import h2o
        from h2o.estimators.glm import H2OGeneralizedLinearEstimator
        h2o.init()

        # import the airlines dataset:
        # This dataset is used to classify whether a flight will be delayed 'YES' or not "NO"
        # original data can be found at http://www.transtats.bts.gov/
        airlines= h2o.import_file("https://s3.amazonaws.com/h2o-public-test-data/smalldata/airlines/allyears2k_headers.zip")

        # convert columns to factors
        airlines["Year"]= airlines["Year"].asfactor()
        airlines["Month"]= airlines["Month"].asfactor()
        airlines["DayOfWeek"] = airlines["DayOfWeek"].asfactor()
        airlines["Cancelled"] = airlines["Cancelled"].asfactor()
        airlines['FlightNum'] = airlines['FlightNum'].asfactor()

        # set the predictor names and the response column name
        predictors = ["Origin", "Dest", "Year", "UniqueCarrier", "DayOfWeek", "Month", "Distance", "FlightNum"]
        response = "IsDepDelayed"

        # split into train and validation sets
        train, valid= airlines.split_frame(ratios = [.8])

        # try using the `control_variables` parameter:
        # initialize your estimator
        airlines_glm = H2OGeneralizedLinearEstimator(family = 'binomial',
                                                     remove_collinear_columns = True,
                                                                                                 score_each_iteration = True,
                                                                                                 generate_scoring_history = True,
                                                     control_variables = ["Year", "DayOfWeek"])

        # then train your model
        airlines_glm.train(x = predictors, y = response, training_frame = train, validation_frame = valid)

        # print the auc for the validation data
        print(airlines_glm.auc(valid=True))

        # take a look at the coefficients_table
        coeff_table = airlines_glm._model_json['output']['coefficients_table']

        # convert table to a pandas dataframe
        coeff_table.as_data_frame()

        # take a look at the learning curve
        airlines_glm.learning_curve_plot()

        # get the unrestricted GLM model
        unrestricted_airlines_glm = airlines_glm.make_unrestricted_glm_model()

# get variable importance tables
varimp = airlines_glm.varimp()
varimp_unrestricted = unrestricted_airlines_glm.varimp()