sample_size

  • Available in: Isolation Forest

  • Hyperparameter: yes

Description

This option specifies the number of randomly sampled observations used to train each Isolation Forest tree. If set to -1, sample_rate will be used instead. This value defaults to 256.

Example

R

Python

library(h2o)
h2o.init()
# import the ecg discord datasets:
train <- h2o.importFile("http://s3.amazonaws.com/h2o-public-test-data/smalldata/anomaly/ecg_discord_train.csv")
test <- h2o.importFile("http://s3.amazonaws.com/h2o-public-test-data/smalldata/anomaly/ecg_discord_test.csv")

# train using the `sample_size` parameter:
isofor_model <- h2o.isolationForest(training_frame=train, sample_size=5, ntrees=7)

# test the prediction
anomaly_score <- h2o.predict(isofor_model, test)
anomaly_score
      predict mean_length
1 -0.16666667    2.857143
2 -0.16666667    2.857143
3 -0.08333333    2.714286
4  0.16666667    2.285714
5  0.00000000    2.571429
6  0.33333333    2.000000

[23 rows x 2 columns]

import h2o
from h2o.estimators.isolation_forest import H2OIsolationForestEstimator
h2o.init()

# import the ecg discord datasets:
train = h2o.import_file("http://s3.amazonaws.com/h2o-public-test-data/smalldata/anomaly/ecg_discord_train.csv")
test = h2o.import_file("http://s3.amazonaws.com/h2o-public-test-data/smalldata/anomaly/ecg_discord_test.csv")

# try using the `sample_size` parameter:
isofor_model = H2OIsolationForestEstimator(sample_size = 5, ntrees=7)

# then train your model
isofor_model.train(training_frame = train)

perf = isofor_model.model_performance()
perf
ModelMetricsAnomaly: isolationforest
** Reported on train data. **

Anomaly Score: 1.1392230576441102
Normalized Anomaly Score: 0.3631578947368421

test_pred = isofor_model.predict(test)
test_pred
predict    mean_length
---------  -------------
-0.1             1.57143
-0.1             1.57143
 0               1.42857
 0               1.42857
 0               1.42857
-0.1             1.57143
 0.1             1.28571
 0               1.42857
 0.2             1.14286
-0.1             1.57143

[23 rows x 2 columns]