Add ConfidenceIntervalOutliersTransform (#196)

* Add ConfidenceIntervalOutliersTransform

* Fixes

* Fix tests

* Update CHANGELOG

* Update Outliers notebook

CHANGELOG.md

@@ -34,6 +34,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - Examples for CatBoostModelMultiSegment and CatBoostModelPerSegment ([#181](https://github.com/tinkoff-ai/etna-ts/pull/181))
 - Simplify TSDataset.train_test_split method by allowing to pass not all values ([#191](https://github.com/tinkoff-ai/etna-ts/pull/191))
 - Confidence interval anomalies detection to EDA ([#182](https://github.com/tinkoff-ai/etna-ts/pull/182))
+- ConfidenceIntervalOutliersTransform ([#196](https://github.com/tinkoff-ai/etna-ts/pull/196))
 
 ### Changed
 - Delete offset from WindowStatisticsTransform ([#111](https://github.com/tinkoff-ai/etna-ts/pull/111))

README.md

@@ -89,6 +89,7 @@ We have also prepared a set of tutorials for an easy introduction:
 - Point outliers
     - Median method
     - Density method
+    - Confidence interval method
 - Sequence outliers
 - Interactive visualization
 - Outliers imputation

etna/transforms/__init__.py

@@ -9,6 +9,7 @@
 from etna.transforms.imputation import TimeSeriesImputerTransform
 from etna.transforms.lags import LagTransform
 from etna.transforms.log import LogTransform
+from etna.transforms.outliers import ConfidenceIntervalOutliersTransform
 from etna.transforms.outliers import DensityOutliersTransform
 from etna.transforms.outliers import MedianOutliersTransform
 from etna.transforms.outliers import SAXOutliersTransform

etna/transforms/outliers.py

@@ -4,14 +4,19 @@
 from typing import Dict
 from typing import List
 from typing import Optional
+from typing import Type
+from typing import Union
 
 import numpy as np
 import pandas as pd
 
+from etna.analysis import get_anomalies_confidence_interval
 from etna.analysis import get_anomalies_density
 from etna.analysis import get_anomalies_median
 from etna.analysis import get_sequence_anomalies
 from etna.datasets import TSDataset
+from etna.models import ProphetModel
+from etna.models import SARIMAXModel
 from etna.transforms.base import Transform
 
 
@@ -265,4 +270,52 @@ def detect_outliers(self, ts: TSDataset) -> Dict[str, List[pd.Timestamp]]:
         )
 
 
-__all__ = ["MedianOutliersTransform", "DensityOutliersTransform"]
+class ConfidenceIntervalOutliersTransform(OutliersTransform):
+    """Transform that uses get_anomalies_density to find anomalies in data."""
+
+    def __init__(
+        self,
+        in_column: str,
+        model: Union[Type["ProphetModel"], Type["SARIMAXModel"]],
+        interval_width: float = 0.95,
+        **model_kwargs,
+    ):
+        """Create instance of ConfidenceIntervalOutliersTransform.
+
+        Parameters
+        ----------
+        ts:
+            TSDataset with timeseries data(should contains all the necessary features)
+        model:
+            model for confidence interval estimation
+        interval_width:
+            width of the confidence interval
+        """
+        self.in_column = in_column
+        self.model = model
+        self.interval_width = interval_width
+        self.model_kwargs = model_kwargs
+        super().__init__(in_column=self.in_column)
+
+    def detect_outliers(self, ts: TSDataset) -> Dict[str, List[pd.Timestamp]]:
+        """Call `get_anomalies_confidence_interval` function with self parameters.
+
+        Parameters
+        ----------
+        ts:
+            dataset to process
+
+        Returns
+        -------
+        dict of outliers:
+            dict of outliers in format {segment: [outliers_timestamps]}
+        """
+        return get_anomalies_confidence_interval(ts, self.model, self.interval_width, **self.model_kwargs)
+
+
+__all__ = [
+    "MedianOutliersTransform",
+    "DensityOutliersTransform",
+    "SAXOutliersTransform",
+    "ConfidenceIntervalOutliersTransform",
+]

tests/test_transforms/test_outliers_transform.py

@@ -2,10 +2,13 @@
 import pandas as pd
 import pytest
 
+from etna.analysis import get_anomalies_confidence_interval
 from etna.analysis import get_anomalies_density
 from etna.analysis import get_anomalies_median
 from etna.analysis import get_sequence_anomalies
 from etna.datasets.tsdataset import TSDataset
+from etna.models import ProphetModel
+from etna.transforms import ConfidenceIntervalOutliersTransform
 from etna.transforms import DensityOutliersTransform
 from etna.transforms import MedianOutliersTransform
 from etna.transforms import SAXOutliersTransform
@@ -43,6 +46,7 @@ def outliers_solid_tsds():
         MedianOutliersTransform(in_column="target"),
         DensityOutliersTransform(in_column="target"),
         SAXOutliersTransform(in_column="target"),
+        ConfidenceIntervalOutliersTransform(in_column="target", model=ProphetModel),
     ],
 )
 def test_interface(transform, example_tsds: TSDataset):
@@ -53,16 +57,21 @@ def test_interface(transform, example_tsds: TSDataset):
 
 
 @pytest.mark.parametrize(
-    "transform, method",
+    "transform, method, method_kwargs",
     [
-        (MedianOutliersTransform(in_column="target"), get_anomalies_median),
-        (DensityOutliersTransform(in_column="target"), get_anomalies_density),
-        (SAXOutliersTransform(in_column="target"), get_sequence_anomalies),
+        (MedianOutliersTransform(in_column="target"), get_anomalies_median, {}),
+        (DensityOutliersTransform(in_column="target"), get_anomalies_density, {}),
+        (SAXOutliersTransform(in_column="target"), get_sequence_anomalies, {}),
+        (
+            ConfidenceIntervalOutliersTransform(in_column="target", model=ProphetModel),
+            get_anomalies_confidence_interval,
+            {"model": ProphetModel},
+        ),
     ],
 )
-def test_outliers_detection(transform, method, outliers_tsds):
+def test_outliers_detection(transform, method, outliers_tsds, method_kwargs):
     """Checks that outliers transforms detect anomalies according to methods from etna.analysis."""
-    detectiom_method_results = method(outliers_tsds)
+    detectiom_method_results = method(outliers_tsds, **method_kwargs)
 
     # save for each segment index without existing nans
     non_nan_index = {}
@@ -82,6 +91,7 @@ def test_outliers_detection(transform, method, outliers_tsds):
         MedianOutliersTransform(in_column="target"),
         DensityOutliersTransform(in_column="target"),
         SAXOutliersTransform(in_column="target"),
+        ConfidenceIntervalOutliersTransform(in_column="target", model=ProphetModel),
     ],
 )
 def test_inverse_transform_train(transform, outliers_solid_tsds):
@@ -99,6 +109,7 @@ def test_inverse_transform_train(transform, outliers_solid_tsds):
         MedianOutliersTransform(in_column="target"),
         DensityOutliersTransform(in_column="target"),
         SAXOutliersTransform(in_column="target"),
+        ConfidenceIntervalOutliersTransform(in_column="target", model=ProphetModel),
     ],
 )
 def test_inverse_transform_future(transform, outliers_solid_tsds):