Makes the reg_covar parameter settable to a value different from the default.

raj-sinha · The spade_anomaly_detection Authors · commit b02af20e7dd3 · 2025-07-07T14:31:15.000-07:00
The new parameter remains an internal parameter for now and is changed only for a unit test.

PiperOrigin-RevId: 780262473
diff --git a/spade_anomaly_detection/__init__.py b/spade_anomaly_detection/__init__.py
@@ -1,4 +1,4 @@
-# Copyright 2024 The spade_anomaly_detection Authors.
+# Copyright 2025 The spade_anomaly_detection Authors.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
diff --git a/spade_anomaly_detection/csv_data_loader.py b/spade_anomaly_detection/csv_data_loader.py
@@ -1,4 +1,4 @@
-# Copyright 2024 The spade_anomaly_detection Authors.
+# Copyright 2025 The spade_anomaly_detection Authors.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
diff --git a/spade_anomaly_detection/csv_data_loader_test.py b/spade_anomaly_detection/csv_data_loader_test.py
@@ -1,4 +1,4 @@
-# Copyright 2024 The spade_anomaly_detection Authors.
+# Copyright 2025 The spade_anomaly_detection Authors.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
diff --git a/spade_anomaly_detection/data_loader.py b/spade_anomaly_detection/data_loader.py
@@ -1,4 +1,4 @@
-# Copyright 2024 The spade_anomaly_detection Authors.
+# Copyright 2025 The spade_anomaly_detection Authors.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
diff --git a/spade_anomaly_detection/data_loader_test.py b/spade_anomaly_detection/data_loader_test.py
@@ -1,4 +1,4 @@
-# Copyright 2024 The spade_anomaly_detection Authors.
+# Copyright 2025 The spade_anomaly_detection Authors.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
diff --git a/spade_anomaly_detection/data_utils/__init__.py b/spade_anomaly_detection/data_utils/__init__.py
@@ -1,4 +1,4 @@
-# Copyright 2024 The spade_anomaly_detection Authors.
+# Copyright 2025 The spade_anomaly_detection Authors.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
diff --git a/spade_anomaly_detection/data_utils/bq_dataset.py b/spade_anomaly_detection/data_utils/bq_dataset.py
@@ -1,4 +1,4 @@
-# Copyright 2024 The spade_anomaly_detection Authors.
+# Copyright 2025 The spade_anomaly_detection Authors.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
diff --git a/spade_anomaly_detection/data_utils/bq_dataset_test.py b/spade_anomaly_detection/data_utils/bq_dataset_test.py
@@ -1,4 +1,4 @@
-# Copyright 2024 The spade_anomaly_detection Authors.
+# Copyright 2025 The spade_anomaly_detection Authors.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
diff --git a/spade_anomaly_detection/data_utils/bq_dataset_test_utils.py b/spade_anomaly_detection/data_utils/bq_dataset_test_utils.py
@@ -1,4 +1,4 @@
-# Copyright 2024 The spade_anomaly_detection Authors.
+# Copyright 2025 The spade_anomaly_detection Authors.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
diff --git a/spade_anomaly_detection/data_utils/bq_utils.py b/spade_anomaly_detection/data_utils/bq_utils.py
@@ -1,4 +1,4 @@
-# Copyright 2024 The spade_anomaly_detection Authors.
+# Copyright 2025 The spade_anomaly_detection Authors.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
diff --git a/spade_anomaly_detection/data_utils/bq_utils_test.py b/spade_anomaly_detection/data_utils/bq_utils_test.py
@@ -1,4 +1,4 @@
-# Copyright 2024 The spade_anomaly_detection Authors.
+# Copyright 2025 The spade_anomaly_detection Authors.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
diff --git a/spade_anomaly_detection/data_utils/feature_metadata.py b/spade_anomaly_detection/data_utils/feature_metadata.py
@@ -1,4 +1,4 @@
-# Copyright 2024 The spade_anomaly_detection Authors.
+# Copyright 2025 The spade_anomaly_detection Authors.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
diff --git a/spade_anomaly_detection/data_utils/feature_metadata_test.py b/spade_anomaly_detection/data_utils/feature_metadata_test.py
@@ -1,4 +1,4 @@
-# Copyright 2024 The spade_anomaly_detection Authors.
+# Copyright 2025 The spade_anomaly_detection Authors.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
diff --git a/spade_anomaly_detection/occ_ensemble.py b/spade_anomaly_detection/occ_ensemble.py
@@ -1,4 +1,4 @@
-# Copyright 2024 The spade_anomaly_detection Authors.
+# Copyright 2025 The spade_anomaly_detection Authors.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
@@ -83,6 +83,9 @@ class GmmEnsemble:
       entire training dataset. This will be used to score samples and classify
       them as anomalies. Note that the more models in the ensemble, the harder
       it is to gain consensus.
+    reg_covar: Non-negative regularization added to the diagonal of covariance.
+      Leave at the default value unless the dataset is very small and there are
+      numerical instabilities during fitting.
     positive_threshold: Float between [0, 100] used as the percentile for the
       one class classifier ensemble to label a point as positive. The closer to
       0 this value is set, the less positive data will be labeled. However, we
@@ -118,6 +121,7 @@ def __init__(
       init_params: str = 'kmeans',
       max_iter: int = 100,
       ensemble_count: int = 5,
+      reg_covar: float = 1e-6,
       positive_threshold: float = 1.0,
       negative_threshold: float = 95.0,
       voting_strategy: parameters.VotingStrategy = (
@@ -135,6 +139,7 @@ def __init__(
     self.init_params = init_params
     self.max_iter = max_iter
     self.ensemble_count = ensemble_count
+    self._reg_covar = reg_covar
     self.positive_threshold = positive_threshold
     self.negative_threshold = negative_threshold
     self.voting_strategy = voting_strategy
@@ -169,6 +174,7 @@ def _get_model(self, idx: int) -> mixture.GaussianMixture:
         warm_start=self._warm_start,
         max_iter=self.max_iter,
         random_state=self._random_seed,
+        reg_covar=self._reg_covar,
     )
 
   def _get_filter_by_label_value_func(self, label_column_filter_value: int):
diff --git a/spade_anomaly_detection/occ_ensemble_test.py b/spade_anomaly_detection/occ_ensemble_test.py
@@ -1,4 +1,4 @@
-# Copyright 2024 The spade_anomaly_detection Authors.
+# Copyright 2025 The spade_anomaly_detection Authors.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
@@ -73,6 +73,9 @@ def test_ensemble_training_no_error(
         n_components=n_components,
         ensemble_count=ensemble_count,
         covariance_type=covariance_type,
+        # `reg_covar` is set to a value greater than the default value of 1e-6
+        # to avoid numerical stability issues in the unit test.
+        reg_covar=1e-5,
     )
 
     tf_dataset = data_loader.load_tf_dataset_from_csv(
diff --git a/spade_anomaly_detection/parameters.py b/spade_anomaly_detection/parameters.py
@@ -1,4 +1,4 @@
-# Copyright 2024 The spade_anomaly_detection Authors.
+# Copyright 2025 The spade_anomaly_detection Authors.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
@@ -187,6 +187,9 @@ class RunnerParameters:
     covariance_type: The covariance type to use in the one class classifier
       ensemble. By default, we use 'full' covariance. Note that when there are
       many components, a 'full' covariance matrix may not be suitable.
+    reg_covar: Non-negative regularization added to the diagonal of covariance.
+      Leave at the default value unless the dataset is very small and there are
+      numerical instabilities during fitting.
     random_seed: The random seed to use for all random number generators in the
       algorithm.
     verbose: The amount of console logs to display during training. Use False to
@@ -223,6 +226,10 @@ class RunnerParameters:
   ensemble_count: int = 5
   n_components: tuple[int, ...] = (1,)
   covariance_type: str = 'full'
+  # In normal use, do not change 'reg_covar'. If your dataset is very small,
+  # then setting it to a larger value may improve the numerical stability of
+  # your model fitting.
+  reg_covar: float = 1e-6
   random_seed: int = _RANDOM_SEED
   verbose: bool = False
 
@@ -258,6 +265,9 @@ def __post_init__(self):
           '`labels_are_strings` must be True if `positive_data_value`, '
           '`negative_data_value` and `unlabeled_data_value` are strings.'
       )
+    if self.reg_covar <= 0:
+      raise ValueError('`reg_covar` must be positive.')
+
     # Adjust the labels if needed.
     if not self.labels_are_strings:
       self.positive_data_value = int(self.positive_data_value)
diff --git a/spade_anomaly_detection/parameters_test.py b/spade_anomaly_detection/parameters_test.py
@@ -1,4 +1,4 @@
-# Copyright 2024 The spade_anomaly_detection Authors.
+# Copyright 2025 The spade_anomaly_detection Authors.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
@@ -37,7 +37,7 @@
 
 class ParametersTest(tf.test.TestCase):
 
-  def test_none_required_parameter_raises(self):
+  def test_incorrect_parameter_raises(self):
     with self.subTest(name='two_input_sources_specified'):
       with self.assertRaises(ValueError):
         _ = parameters.RunnerParameters(
@@ -64,6 +64,20 @@ def test_none_required_parameter_raises(self):
             unlabeled_data_value='-1',
             voting_strategy=parameters.VotingStrategy.UNANIMOUS,
         )
+    with self.subTest(name='reg_covar_not_positive'):
+      with self.assertRaises(ValueError):
+        _ = parameters.RunnerParameters(
+            train_setting=parameters.TrainSetting.PNU,
+            input_bigquery_table_path='some_project.some_dataset.some_table',
+            data_input_gcs_uri=None,
+            output_gcs_uri='gs://some_bucket/some_path',
+            label_col_name='y',
+            positive_data_value='1',
+            negative_data_value='0',
+            unlabeled_data_value='-1',
+            voting_strategy=parameters.VotingStrategy.UNANIMOUS,
+            reg_covar=-1e-6,
+        )
 
   def test_equal_data_value_parameter_raises(self):
     with self.assertRaises(ValueError):
diff --git a/spade_anomaly_detection/performance_test.py b/spade_anomaly_detection/performance_test.py
@@ -1,4 +1,4 @@
-# Copyright 2024 The spade_anomaly_detection Authors.
+# Copyright 2025 The spade_anomaly_detection Authors.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
diff --git a/spade_anomaly_detection/runner.py b/spade_anomaly_detection/runner.py
@@ -1,4 +1,4 @@
-# Copyright 2024 The spade_anomaly_detection Authors.
+# Copyright 2025 The spade_anomaly_detection Authors.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
diff --git a/spade_anomaly_detection/runner_test.py b/spade_anomaly_detection/runner_test.py
@@ -1,4 +1,4 @@
-# Copyright 2024 The spade_anomaly_detection Authors.
+# Copyright 2025 The spade_anomaly_detection Authors.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
diff --git a/spade_anomaly_detection/supervised_model.py b/spade_anomaly_detection/supervised_model.py
@@ -1,4 +1,4 @@
-# Copyright 2024 The spade_anomaly_detection Authors.
+# Copyright 2025 The spade_anomaly_detection Authors.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
diff --git a/spade_anomaly_detection/supervised_model_test.py b/spade_anomaly_detection/supervised_model_test.py
@@ -1,4 +1,4 @@
-# Copyright 2024 The spade_anomaly_detection Authors.
+# Copyright 2025 The spade_anomaly_detection Authors.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
diff --git a/spade_anomaly_detection/task.py b/spade_anomaly_detection/task.py
@@ -1,4 +1,4 @@
-# Copyright 2024 The spade_anomaly_detection Authors.
+# Copyright 2025 The spade_anomaly_detection Authors.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.

Original file line number	Diff line number	Diff line change
`@@ -1,4 +1,4 @@`
`1`		`-# Copyright 2024 The spade_anomaly_detection Authors.`
	`1`	`+# Copyright 2025 The spade_anomaly_detection Authors.`
`2`	`2`	`#`
`3`	`3`	`# Licensed under the Apache License, Version 2.0 (the "License");`
`4`	`4`	`# you may not use this file except in compliance with the License.`