max_models

• Available in: AutoML

• Hyperparameter: no

Description

Use this option to specify the maximum number of models to build in the AutoML run, excluding the Stacked Ensemble models. This option defaults to Null/None. This option should systematically be set if AutoML reproducibility is needed: all models are then trained until convergence and none is constrained by a time budget.

Related Parameters

• max_runtime_secs

• max_runtime_secs_per_model

• seed

Example

R

Python

library(h2o)
h2o.init()

# Import the prostate dataset
prostate <- h2o.importFile("http://s3.amazonaws.com/h2o-public-test-data/smalldata/prostate/prostate_complete.csv.zip")

# Set the predictor names and the response column name
y <- "CAPSULE"
x <- setdiff(names(prostate), c(p_y, "ID"))

# Train AutoML
aml <- h2o.automl(x = x,
                  y = y,
                  training_frame = prostate,
                  seed = 1234,
                  max_models = 5,
                  max_runtime_secs = 200,
                  max_runtime_secs_per_model = 40)

# View the AutoML Leaderboard
lb <- aml@leaderboard
lb

                                                     model_id mean_residual_deviance
1 StackedEnsemble_BestOfFamily_AutoML_20190321_110032           0.0009730593
2    StackedEnsemble_AllModels_AutoML_20190321_110032           0.0009730593
3                        DRF_1_AutoML_20190321_110032           0.0012766064
4                        XRT_1_AutoML_20190321_110032           0.0038347775
5                    XGBoost_2_AutoML_20190321_110032           0.0064206276
6                    XGBoost_1_AutoML_20190321_110032           0.0544174809
        rmse          mse        mae      rmsle
1 0.03119390 0.0009730593 0.02086672 0.02368844
2 0.03119390 0.0009730593 0.02086672 0.02368844
3 0.03572963 0.0012766064 0.01406001 0.02661268
4 0.06192558 0.0038347775 0.03330358 0.04958889
5 0.08012882 0.0064206276 0.06873394 0.06112533
6 0.23327555 0.0544174809 0.18390358 0.16640402

[7 rows x 6 columns]

import h2o
from h2o.automl import H2OAutoML
h2o.init()

# Import a sample binary outcome training set into H2O
prostate = h2o.import_file("http://s3.amazonaws.com/h2o-public-test-data/smalldata/prostate/prostate_complete.csv.zip")

# Set the predictor names and the response column name
response = "CAPSULE"
predictor = prostate.names[2:9]

# Train AutoML
aml = H2OAutoML(max_models = 5,
                max_runtime_secs = 200,
                max_runtime_secs_per_model = 40,
                seed = 1234)
aml.train(x = predictor, y = response, training_frame = prostate)

# View the AutoML Leaderboard
lb = aml.leaderboard
lb
model_id                                               mean_residual_deviance       rmse          mse        mae      rmsle
---------------------------------------------------  ------------------------  ---------  -----------  ---------  ---------
StackedEnsemble_AllModels_AutoML_20190321_111608                  0.000282073  0.016795   0.000282073  0.0103226  0.0129982
StackedEnsemble_BestOfFamily_AutoML_20190321_111608               0.000282073  0.016795   0.000282073  0.0103226  0.0129982
DRF_1_AutoML_20190321_111608                                      0.000334287  0.0182835  0.000334287  0.0076525  0.0140754
XRT_1_AutoML_20190321_111608                                      0.0015397    0.039239   0.0015397    0.0217268  0.0293752
XGBoost_2_AutoML_20190321_111608                                  0.0118094    0.108671   0.0118094    0.0888375  0.0804565
XGBoost_1_AutoML_20190321_111608                                  0.0675672    0.259937   0.0675672    0.213536   0.184793
GLM_grid_1_AutoML_20190321_111608_model_1                         0.193551     0.439944   0.193551     0.397327   0.306996

[7 rows x 6 columns]