Merge pull request #224 from IFCA/fix-include-cusum-change-detection

Jaime Céspedes Sisniega · web-flow · commit cea5af493d5f · 2023-06-21T13:42:25.000+02:00
Include CUSUM methods into change detection
diff --git a/README.md b/README.md
@@ -209,14 +209,13 @@ The currently implemented detectors are listed in the following table.
   <tr>
     <td rowspan="13" style="text-align: center; border: 1px solid grey; padding: 8px;">Concept drift</td>
     <td rowspan="13" style="text-align: center; border: 1px solid grey; padding: 8px;">Streaming</td>
-    <td rowspan="1" style="text-align: center; border: 1px solid grey; padding: 8px;">Change detection</td>
+    <td rowspan="4" style="text-align: center; border: 1px solid grey; padding: 8px;">Change detection</td>
     <td style="text-align: center; border: 1px solid grey; padding: 8px;">U</td>
     <td style="text-align: center; border: 1px solid grey; padding: 8px;">N</td>
     <td style="text-align: center; border: 1px solid grey; padding: 8px;">BOCD</td>
     <td style="text-align: center; border: 1px solid grey; padding: 8px;"><a href="https://doi.org/10.48550/arXiv.0710.3742">Adams and MacKay (2007)</a></td>
   </tr>
   <tr>
-    <td rowspan="3" style="text-align: center; border: 1px solid grey; padding: 8px;">CUSUM</td>
     <td style="text-align: center; border: 1px solid grey; padding: 8px;">U</td>
     <td style="text-align: center; border: 1px solid grey; padding: 8px;">N</td>
     <td style="text-align: center; border: 1px solid grey; padding: 8px;">CUSUM</td>
diff --git a/docs/source/api_reference/detectors/concept_drift/streaming.md b/docs/source/api_reference/detectors/concept_drift/streaming.md
@@ -26,21 +26,6 @@ The {mod}`frouros.detectors.concept_drift.streaming` module contains streaming c
 
     BOCD
     BOCDConfig
-```
-
-## CUSUM Test
-
-```{eval-rst}
-.. automodule:: frouros.detectors.concept_drift.streaming.cusum_based
-    :no-members:
-    :no-inherited-members:
-```
-
-```{eval-rst}
-.. autosummary::
-    :toctree: auto_generated/
-    :template: class.md
-
     CUSUM
     CUSUMConfig
     GeometricMovingAverage
diff --git a/frouros/detectors/concept_drift/streaming/__init__.py b/frouros/detectors/concept_drift/streaming/__init__.py
@@ -4,9 +4,6 @@
 from .change_detection import (
     BOCD,
     BOCDConfig,
-)
-
-from .cusum_based import (
     CUSUM,
     CUSUMConfig,
     GeometricMovingAverage,
diff --git a/frouros/detectors/concept_drift/streaming/change_detection/__init__.py b/frouros/detectors/concept_drift/streaming/change_detection/__init__.py
@@ -1,8 +1,29 @@
 """Concept drift change detection methods' init."""
 
-from .bocd import BOCD, BOCDConfig
+from .bocd import (
+    BOCD,
+    BOCDConfig,
+)
+from .cusum import (
+    CUSUM,
+    CUSUMConfig,
+)
+from .geometric_moving_average import (
+    GeometricMovingAverage,
+    GeometricMovingAverageConfig,
+)
+from .page_hinkley import (
+    PageHinkley,
+    PageHinkleyConfig,
+)
 
 __all__ = [
     "BOCD",
     "BOCDConfig",
+    "CUSUM",
+    "CUSUMConfig",
+    "GeometricMovingAverage",
+    "GeometricMovingAverageConfig",
+    "PageHinkley",
+    "PageHinkleyConfig",
 ]
diff --git a/frouros/detectors/concept_drift/streaming/change_detection/base.py b/frouros/detectors/concept_drift/streaming/change_detection/base.py
@@ -1,12 +1,14 @@
 """Base concept drift ChangeDetection based module."""
 
 import abc
-from typing import Union
+from typing import List, Optional, Union
 
+from frouros.callbacks.streaming.base import BaseCallbackStreaming
 from frouros.detectors.concept_drift.streaming.base import (
     BaseConceptDriftStreaming,
     BaseConceptDriftStreamingConfig,
 )
+from frouros.utils.stats import Mean
 
 
 class BaseChangeDetectionConfig(BaseConceptDriftStreamingConfig):
@@ -21,3 +23,207 @@ class BaseChangeDetection(BaseConceptDriftStreaming):
     @abc.abstractmethod
     def _update(self, value: Union[int, float], **kwargs) -> None:
         pass
+
+
+class BaseCUSUMConfig(BaseChangeDetectionConfig):
+    """Class representing a CUSUM based configuration class."""
+
+    def __init__(
+        self,
+        lambda_: float = 50.0,
+        min_num_instances: int = 30,
+    ) -> None:
+        """Init method.
+
+        :param lambda_: lambda value
+        :type lambda_: float
+        :param min_num_instances: minimum numbers of instances
+        to start looking for changes
+        :type min_num_instances: int
+        """
+        super().__init__(min_num_instances=min_num_instances)
+        self.lambda_ = lambda_
+
+    @property
+    def lambda_(self) -> float:
+        """Threshold property.
+
+        :return: lambda to use
+        :rtype: float
+        """
+        return self._lambda
+
+    @lambda_.setter
+    def lambda_(self, value: float) -> None:
+        """Threshold setter.
+
+        :param value: value to be set
+        :type value: float
+        :raises ValueError: Value error exception
+        """
+        if value < 0:
+            raise ValueError("lambda_ must be great or equal than 0.")
+        self._lambda = value
+
+
+class DeltaConfig:
+    """Class representing a delta configuration class."""
+
+    def __init__(
+        self,
+        delta: float = 0.005,
+    ) -> None:
+        """Init method.
+
+        :param delta: delta value
+        :type delta: float
+        """
+        self.delta = delta
+
+    @property
+    def delta(self) -> float:
+        """Delta property.
+
+        :return: delta to use
+        :rtype: float
+        """
+        return self._delta
+
+    @delta.setter
+    def delta(self, value: float) -> None:
+        """Delta setter.
+
+        :param value: value to be set
+        :type value: float
+        :raises ValueError: Value error exception
+        """
+        if not 0.0 <= value <= 1.0:
+            raise ValueError("delta must be in the range [0, 1].")
+        self._delta = value
+
+
+class AlphaConfig:
+    """Class representing an alpha configuration class."""
+
+    def __init__(
+        self,
+        alpha: float = 0.9999,
+    ) -> None:
+        """Init method.
+
+        :param alpha: forgetting factor value
+        :type alpha: float
+        """
+        self.alpha = alpha
+
+    @property
+    def alpha(self) -> float:
+        """Forgetting factor property.
+
+        :return: forgetting factor value
+        :rtype: float
+        """
+        return self._alpha
+
+    @alpha.setter
+    def alpha(self, value: float) -> None:
+        """Forgetting factor setter.
+
+        :param value: forgetting factor value
+        :type value: float
+        :raises ValueError: Value error exception
+        """
+        if not 0.0 <= value <= 1.0:
+            raise ValueError("alpha must be in the range [0, 1].")
+        self._alpha = value
+
+
+class BaseCUSUM(BaseChangeDetection):
+    """CUSUM based algorithm class."""
+
+    config_type = BaseCUSUMConfig
+
+    def __init__(
+        self,
+        config: Optional[BaseCUSUMConfig] = None,
+        callbacks: Optional[
+            Union[BaseCallbackStreaming, List[BaseCallbackStreaming]]
+        ] = None,
+    ) -> None:
+        """Init method.
+
+        :param config: configuration parameters
+        :type config: Optional[BaseCUSUMConfig]
+        :param callbacks: callbacks
+        :type callbacks: Optional[Union[BaseCallbackStreaming,
+        List[BaseCallbackStreaming]]]
+        """
+        super().__init__(
+            config=config,
+            callbacks=callbacks,
+        )
+        self.additional_vars = {
+            "mean_error_rate": Mean(),
+            "sum_": 0.0,
+        }
+        self._set_additional_vars_callback()
+
+    @property
+    def mean_error_rate(self) -> Mean:
+        """Mean error rate property.
+
+        :return: mean error rate to use
+        :rtype: Mean
+        """
+        return self._additional_vars["mean_error_rate"]
+
+    @mean_error_rate.setter
+    def mean_error_rate(self, value: Mean) -> None:
+        """Mean error rate setter.
+
+        :param value: value to be set
+        :type value: Mean
+        """
+        self._additional_vars["mean_error_rate"] = value
+
+    @property
+    def sum_(self) -> float:
+        """Sum count property.
+
+        :return: sum count value
+        :rtype: float
+        """
+        return self._additional_vars["sum_"]
+
+    @sum_.setter
+    def sum_(self, value: float) -> None:
+        """Sum count setter.
+
+        :param value: value to be set
+        :type value: float
+        """
+        self._additional_vars["sum_"] = value
+
+    @abc.abstractmethod
+    def _update_sum(self, error_rate: float) -> None:
+        pass
+
+    def reset(self) -> None:
+        """Reset method."""
+        super().reset()
+        self.mean_error_rate = Mean()
+        self.sum_ = 0.0
+
+    def _update(self, value: Union[int, float], **kwargs) -> None:
+        self.num_instances += 1
+
+        self.mean_error_rate.update(value=value)
+        self._update_sum(error_rate=value)
+
+        if (
+            self.num_instances >= self.config.min_num_instances  # type: ignore
+            and self.sum_ > self.config.lambda_  # type: ignore
+        ):
+            self.drift = True
+        else:
+            self.drift = False
diff --git a/frouros/detectors/concept_drift/streaming/change_detection/bocd.py b/frouros/detectors/concept_drift/streaming/change_detection/bocd.py
@@ -47,7 +47,7 @@ class GaussianUnknownMean(BaseBOCDModel):
     def __init__(
         self,
         prior_mean: float = 0,
-        prior_var: float = 0,
+        prior_var: float = 1,
         data_var: float = 1,
     ) -> None:
         """Init method.
diff --git a/frouros/detectors/concept_drift/streaming/change_detection/cusum.py b/frouros/detectors/concept_drift/streaming/change_detection/cusum.py
@@ -2,7 +2,7 @@
 
 import numpy as np  # type: ignore
 
-from frouros.detectors.concept_drift.streaming.cusum_based.base import (
+from frouros.detectors.concept_drift.streaming.change_detection.base import (
     BaseCUSUM,
     BaseCUSUMConfig,
     DeltaConfig,
diff --git a/frouros/detectors/concept_drift/streaming/change_detection/geometric_moving_average.py b/frouros/detectors/concept_drift/streaming/change_detection/geometric_moving_average.py
@@ -1,6 +1,6 @@
 """Geometric Moving Average module."""
 
-from frouros.detectors.concept_drift.streaming.cusum_based.base import (
+from frouros.detectors.concept_drift.streaming.change_detection.base import (
     BaseCUSUM,
     BaseCUSUMConfig,
     AlphaConfig,
diff --git a/frouros/detectors/concept_drift/streaming/change_detection/page_hinkley.py b/frouros/detectors/concept_drift/streaming/change_detection/page_hinkley.py
@@ -1,6 +1,6 @@
 """Page Hinkley module."""
 
-from frouros.detectors.concept_drift.streaming.cusum_based.base import (
+from frouros.detectors.concept_drift.streaming.change_detection.base import (
     BaseCUSUM,
     BaseCUSUMConfig,
     DeltaConfig,
diff --git a/frouros/detectors/concept_drift/streaming/cusum_based/__init__.py b/frouros/detectors/concept_drift/streaming/cusum_based/__init__.py
diff --git a/frouros/detectors/concept_drift/streaming/cusum_based/base.py b/frouros/detectors/concept_drift/streaming/cusum_based/base.py
diff --git a/frouros/tests/integration/test_change_detection.py b/frouros/tests/integration/test_change_detection.py
diff --git a/frouros/tests/integration/test_concept_drift.py b/frouros/tests/integration/test_concept_drift.py
