Populate base_models attribute in Stacked Ensemble in Python and R

[exalate-issue-sync]

Example showing that it's blank:

{code:python}import h2o
from h2o.automl import H2OAutoML
h2o.init()

Import a sample binary outcome train/test set into H2O

train = h2o.import_file("https://s3.amazonaws.com/erin-data/higgs/higgs_train_10k.csv")
test = h2o.import_file("https://s3.amazonaws.com/erin-data/higgs/higgs_test_5k.csv")

Identify predictors and response

x = train.columns
y = "response"
x.remove(y)

For binary classification, response should be a factor

train[y] = train[y].asfactor()
test[y] = test[y].asfactor()
aml = H2OAutoML(max_models=5, seed=1)
aml.train(x=x, y=y, training_frame=train)

In [12]: aml.leader.algo
Out[12]: 'stackedensemble'
In [11]: aml.leader.base_models #this is blank...{code}

The leader model is a stacked ensemble.

There is a dictionary of info about the base models here, aml.leader.get_params()\['base_models'], but i think it would be more useful to return a list of the actual model IDs instead of the dictionary with metadata that appears there. e.g. {{base_models = ["GBM_model_R_1582581833131_73", "DRF_model_R_1582581833131_201"]}}

In R, we don’t currently have an attribute for {{base_models}} so we will need to add one (similar to how we store extra info in the {{metalearner}} attribute).

[h2o-ops]

JIRA Issue Migration Info

Jira Issue: PUBDEV-6787
Assignee: Tomas Fryda
Reporter: Erin LeDell
State: Resolved
Fix Version: 3.28.1.2
Attachments: N/A
Development PRs: Available

Linked PRs from JIRA

#4388