ensemble {autoEnsemble}

Evaluate H2O Model(s) Performance

Description

Multiple model performance metrics are computed

Usage

ensemble(
  models,
  training_frame,
  newdata = NULL,
  family = "binary",
  strategy = c("search"),
  model_selection_criteria = c("auc", "aucpr", "mcc", "f2"),
  min_improvement = 1e-05,
  max = NULL,
  top_rank = seq(0.01, 0.99, 0.01),
  stop_rounds = 3,
  reset_stop_rounds = TRUE,
  stop_metric = "auc",
  seed = -1,
  verbatim = FALSE
)

Arguments

models

H2O search grid or AutoML grid or a character vector of H2O model IDs. the "h2o.get_ids" function from "h2otools" can retrieve the IDs from grids.

training_frame

h2o training frame (data.frame) for model training

newdata

h2o frame (data.frame). the data.frame must be already uploaded on h2o server (cloud). when specified, this dataset will be used for evaluating the models. if not specified, model performance on the training dataset will be reported.

family

model family. currently only "binary" classification models are supported.

strategy

character. the current available strategies are "search" (default) and "top". The "search" strategy searches for the best combination of top-performing diverse models whereas the "top" strategy is more simplified and just combines the specified of top-performing diverse models without examining the possibility of improving the model by searching for larger number of models that can further improve the model. generally, the "search" strategy is preferable, unless the computation runtime is too large and optimization is not possible.

model_selection_criteria

character, specifying the performance metrics that should be taken into consideration for model selection. the default are "c('auc', 'aucpr', 'mcc', 'f2')". other possible criteria are "'f1point5', 'f3', 'f4', 'f5', 'kappa', 'mean_per_class_error', 'gini', 'accuracy'", which are also provided by the "evaluate" function.

min_improvement

numeric. specifies the minimum improvement in model evaluation metric to qualify further optimization search.

max

integer. specifies maximum number of models for each criteria to be extracted. the default value is the "top_rank" percentage for each model selection criteria.

top_rank

numeric vector. specifies percentage of the top models taht should be selected. if the strategy is "search", the algorithm searches for the best best combination of the models from top ranked models to the bottom. however, if the strategy is "top", only the first value of the vector is used (default value is top 1%).

stop_rounds

integer. number of stoping rounds, in case the model stops improving

reset_stop_rounds

logical. if TRUE, everytime the model improves the stopping rounds penalty is resets to 0.

stop_metric

character. model stopping metric. the default is "auc", but "aucpr" and "mcc" are also available.

seed

random seed (recommended)

verbatim

logical. if TRUE, it reports additional information about the progress of the model training, particularly used for debugging.

Value

a matrix of F-Measures for different thresholds or the highest F-Measure value

Author(s)

E. F. Haghish

Examples

## Not run: 
# load the required libraries for building the base-learners and the ensemble models
library(h2o)
library(autoEnsemble)

# initiate the h2o server
h2o.init(ignore_config = TRUE, nthreads = 2, bind_to_localhost = FALSE, insecure = TRUE)

# upload data to h2o cloud
prostate_path <- system.file("extdata", "prostate.csv", package = "h2o")
prostate <- h2o.importFile(path = prostate_path, header = TRUE)

### H2O provides 2 types of grid search for tuning the models, which are
### AutoML and Grid. Below, I tune 2 set of model grids and use them both
### for building the ensemble, just to set an example ...

#######################################################
### PREPARE AutoML Grid (takes a couple of minutes)
#######################################################
# run AutoML to tune various models (GLM, GBM, XGBoost, DRF, DeepLearning) for 120 seconds
y <- "CAPSULE"
prostate[,y] <- as.factor(prostate[,y])  #convert to factor for classification
aml <- h2o.automl(y = y, training_frame = prostate, max_runtime_secs = 120,
                 include_algos=c("DRF","GLM", "XGBoost", "GBM", "DeepLearning"),

                 # this setting ensures the models are comparable for building a meta learner
                 seed = 2023, nfolds = 10,
                 keep_cross_validation_predictions = TRUE)

#######################################################
### PREPARE H2O Grid (takes a couple of minutes)
#######################################################
# make sure equal number of "nfolds" is specified for different grids
grid <- h2o.grid(algorithm = "gbm", y = y, training_frame = prostate,
                 hyper_params = list(ntrees = seq(1,50,1)),
                 grid_id = "ensemble_grid",

                 # this setting ensures the models are comparable for building a meta learner
                 seed = 2023, fold_assignment = "Modulo", nfolds = 10,
                 keep_cross_validation_predictions = TRUE)

#######################################################
### PREPARE ENSEMBLE MODEL
#######################################################

### get the models' IDs from the AutoML and grid searches.
### this is all that is needed before building the ensemble,
### i.e., to specify the model IDs that should be evaluated.

ids    <- c(h2o.get_ids(aml), h2o.get_ids(grid))
top    <- ensemble(models = ids, training_frame = prostate, strategy = "top")
search <- ensemble(models = ids, training_frame = prostate, strategy = "search")

#######################################################
### EVALUATE THE MODELS
#######################################################
h2o.auc(aml@leader)                          # best model identified by h2o.automl
h2o.auc(h2o.getModel(grid@model_ids[[1]]))   # best model identified by grid search
h2o.auc(top).                                # ensemble model with 'top' search strategy
h2o.auc(search).                             # ensemble model with 'search' search strategy


## End(Not run)