dataeval 0.71.1__py3-none-any.whl → 0.72.1__py3-none-any.whl

dataeval/__init__.py

@@ -1,4 +1,4 @@
-__version__ = "0.71.1"
+__version__ = "0.72.1"
 
 from importlib.util import find_spec
 

dataeval/_internal/datasets.py

@@ -185,7 +185,7 @@ class MNIST(Dataset):
         unit_interval : bool, default False
             Shift the data values to the unit interval [0-1].
         dtype : type | None, default None
-            Change the numpy dtype - data is loaded as np.uint8
+            Change the :term:`NumPy` dtype - data is loaded as np.uint8
         channels : Literal['channels_first' | 'channels_last'] | None, default None
             Location of channel axis if desired, default has no channels (N, 28, 28)
         flatten : bool, default False

dataeval/_internal/detectors/clusterer.py

@@ -25,7 +25,7 @@ class ClustererOutput(OutputMetadata):
     potential_outliers : List[int]
         Indices which are near the border between belonging in the cluster and being an outlier
     duplicates : List[List[int]]
-        Groups of indices that are exact duplicates
+        Groups of indices that are exact :term:`duplicates<Duplicates>`
     potential_duplicates : List[List[int]]
         Groups of indices which are not exact but closely related data points
     """
@@ -129,7 +129,7 @@ class ClusterMergeEntry:
 
 class Clusterer:
     """
-    Uses hierarchical clustering to flag dataset properties of interest like outliers and duplicates
+    Uses hierarchical clustering to flag dataset properties of interest like Outliers and :term:`duplicates<Duplicates>`
 
     Parameters
     ----------
@@ -401,7 +401,7 @@ class Clusterer:
 
     def find_outliers(self, last_merge_levels: dict[int, int]) -> tuple[list[int], list[int]]:
         """
-        Retrieves outliers based on when the sample was added to the cluster
+        Retrieves Outliers based on when the sample was added to the cluster
         and how far it was from the cluster when it was added
 
         Parameters
@@ -470,7 +470,7 @@ class Clusterer:
         Returns
         -------
         Tuple[List[List[int]], List[List[int]]]
-            The exact duplicates and near duplicates as lists of related indices
+            The exact :term:`duplicates<Duplicates>` and near duplicates as lists of related indices
         """
 
         duplicates_std = []
@@ -495,12 +495,12 @@ class Clusterer:
     # TODO: Move data input to evaluate from class
     @set_metadata("dataeval.detectors", ["data"])
     def evaluate(self) -> ClustererOutput:
-        """Finds and flags indices of the data for outliers and duplicates
+        """Finds and flags indices of the data for Outliers and :term:`duplicates<Duplicates>`
 
         Returns
         -------
         ClustererOutput
-            The outliers and duplicate indices found in the data
+            The Outliers and duplicate indices found in the data
 
         Example
         -------

dataeval/_internal/detectors/drift/base.py

@@ -47,7 +47,7 @@ class DriftOutput(DriftBaseOutput):
     Attributes
     ----------
     is_drift : bool
-        Drift prediction for the images
+        :term:`Drift` prediction for the images
     threshold : float
         Threshold after multivariate correction if needed
     feature_drift : NDArray
@@ -103,7 +103,7 @@ def preprocess_x(fn):
 
 class UpdateStrategy(ABC):
     """
-    Updates reference dataset for drift detector
+    Updates reference dataset for :term:`drift<Drift>` detector
 
     Parameters
     ----------
@@ -121,7 +121,7 @@ class UpdateStrategy(ABC):
 
 class LastSeenUpdate(UpdateStrategy):
     """
-    Updates reference dataset for drift detector using last seen method.
+    Updates reference dataset for :term:`drift<Drift>` detector using last seen method.
 
     Parameters
     ----------
@@ -136,7 +136,7 @@ class LastSeenUpdate(UpdateStrategy):
 
 class ReservoirSamplingUpdate(UpdateStrategy):
     """
-    Updates reference dataset for drift detector using reservoir sampling method.
+    Updates reference dataset for :term:`drift<Drift>` detector using reservoir sampling method.
 
     Parameters
     ----------
@@ -167,7 +167,7 @@ class ReservoirSamplingUpdate(UpdateStrategy):
 
 class BaseDrift:
     """
-    A generic drift detection component for preprocessing data and applying statistical correction.
+    A generic :term:`drift<Drift>` detection component for preprocessing data and applying statistical correction.
 
     This class handles common tasks related to drift detection, such as preprocessing
     the reference data (`x_ref`), performing statistical correction (e.g., Bonferroni, FDR),
@@ -266,7 +266,7 @@ class BaseDrift:
 
     def _preprocess(self, x: ArrayLike) -> ArrayLike:
         """
-        Preprocess the given data before computing the drift scores.
+        Preprocess the given data before computing the :term:`drift<Drift>` scores.
 
         Parameters
         ----------
@@ -285,12 +285,13 @@ class BaseDrift:
 
 class BaseDriftUnivariate(BaseDrift):
     """
-    Base class for drift detection methods using univariate statistical tests.
+    Base class for :term:`drift<Drift>` detection methods using univariate statistical tests.
 
     This class inherits from `BaseDrift` and serves as a generic component for detecting
     distribution drift in univariate features. If the number of features `n_features` is greater
     than 1, a multivariate correction method (e.g., Bonferroni or FDR) is applied to control
-    the false positive rate, ensuring it does not exceed the specified p-value.
+    the :term:`false positive rate<False Positive Rate (FP)>`, ensuring it does not exceed the specified
+    :term:`p-value<P-Value>`.
 
     Parameters
     ----------
@@ -318,7 +319,7 @@ class BaseDriftUnivariate(BaseDrift):
     p_val : float
         The significance level for drift detection.
     correction : str
-        The method for controlling the false discovery rate or applying a Bonferroni correction.
+        The method for controlling the :term:`False Discovery Rate (FDR)` or applying a Bonferroni correction.
     update_x_ref : UpdateStrategy | None
         Strategy for updating the reference data if applicable.
     preprocess_fn : Callable | None
@@ -393,19 +394,19 @@ class BaseDriftUnivariate(BaseDrift):
         Parameters
         ----------
         x : ArrayLike
-            The batch of data to calculate univariate drift scores for each feature.
+            The batch of data to calculate univariate :term:`drift<Drift>` scores for each feature.
 
         Returns
         -------
         tuple[NDArray, NDArray]
-            A tuple containing p-values and distance statistics for each feature.
+            A tuple containing p-values and distance :term:`statistics<Statistics>` for each feature.
         """
 
     def _apply_correction(self, p_vals: NDArray) -> tuple[bool, float]:
         """
         Apply the specified correction method (Bonferroni or FDR) to the p-values.
 
-        If the correction method is Bonferroni, the threshold for detecting drift
+        If the correction method is Bonferroni, the threshold for detecting :term:`drift<Drift>`
         is divided by the number of features. For FDR, the correction is applied
         using the Benjamini-Hochberg procedure.
 
@@ -457,8 +458,8 @@ class BaseDriftUnivariate(BaseDrift):
         Returns
         -------
         DriftOutput
-            Dictionary containing the drift prediction and optionally the feature level
-            p-values, threshold after multivariate correction if needed and test statistics.
+            Dictionary containing the :term:`drift<Drift>` prediction and optionally the feature level
+            p-values, threshold after multivariate correction if needed and test :term:`statistics<Statistics>`.
         """
         # compute drift scores
         p_vals, dist = self.score(x)

dataeval/_internal/detectors/drift/cvm.py

@@ -21,19 +21,19 @@ from .base import BaseDriftUnivariate, UpdateStrategy, preprocess_x
 
 class DriftCVM(BaseDriftUnivariate):
     """
-    Drift detector employing the Cramér-von Mises (CVM) distribution test.
+    :term:`Drift` detector employing the :term:`Cramér-von Mises (CVM) Drift Detection` test.
 
     The CVM test detects changes in the distribution of continuous
     univariate data. For multivariate data, a separate CVM test is applied to each
     feature, and the obtained p-values are aggregated via the Bonferroni or
-    False Discovery Rate (FDR) corrections.
+    :term:`False Discovery Rate (FDR)` corrections.
 
     Parameters
     ----------
     x_ref : ArrayLike
         Data used as reference distribution.
     p_val : float | None, default 0.05
-        p-value used for significance of the statistical test for each feature.
+        :term:`p-value<P-Value>` used for significance of the statistical test for each feature.
         If the FDR correction method is used, this corresponds to the acceptable
         q-value.
     x_ref_preprocessed : bool, default False
@@ -46,7 +46,7 @@ class DriftCVM(BaseDriftUnivariate):
         or via reservoir sampling with ReservoirSamplingUpdateStrategy.
     preprocess_fn : Callable | None, default None
         Function to preprocess the data before computing the data drift metrics.
-        Typically a dimensionality reduction technique.
+        Typically a :term:`dimensionality reduction<Dimensionality Reduction>` technique.
     correction : "bonferroni" | "fdr", default "bonferroni"
         Correction type for multivariate data. Either 'bonferroni' or 'fdr' (False
         Discovery Rate).
@@ -79,7 +79,7 @@ class DriftCVM(BaseDriftUnivariate):
     @preprocess_x
     def score(self, x: ArrayLike) -> tuple[NDArray[np.float32], NDArray[np.float32]]:
         """
-        Performs the two-sample Cramér-von Mises test(s), computing the p-value and
+        Performs the two-sample Cramér-von Mises test(s), computing the :term:`p-value<P-value>` and
         test statistic per feature.
 
         Parameters

dataeval/_internal/detectors/drift/ks.py

@@ -21,10 +21,10 @@ from .base import BaseDriftUnivariate, UpdateStrategy, preprocess_x
 
 class DriftKS(BaseDriftUnivariate):
     """
-    Drift detector employing the Kolmogorov-Smirnov (KS) distribution test.
+    :term:`Drift` detector employing the Kolmogorov-Smirnov (KS) distribution test.
 
     The KS test detects changes in the maximum distance between two data
-    distributions with Bonferroni or False Discovery Rate (FDR) correction
+    distributions with Bonferroni or :term:`False Discovery Rate (FDR)` correction
     for multivariate data.
 
     Parameters
@@ -32,7 +32,7 @@ class DriftKS(BaseDriftUnivariate):
     x_ref : ArrayLike
         Data used as reference distribution.
     p_val : float | None, default 0.05
-        p-value used for significance of the statistical test for each feature.
+        :term:`p-value<P-Value>` used for significance of the statistical test for each feature.
         If the FDR correction method is used, this corresponds to the acceptable
         q-value.
     x_ref_preprocessed : bool, default False
@@ -44,8 +44,8 @@ class DriftKS(BaseDriftUnivariate):
         using the last n instances seen by the detector with LastSeenUpdateStrategy
         or via reservoir sampling with ReservoirSamplingUpdateStrategy.
     preprocess_fn : Callable | None, default None
-        Function to preprocess the data before computing the data drift metrics.
-        Typically a dimensionality reduction technique.
+        Function to preprocess the data before computing the data :term:`drift<Drift>` metrics.
+        Typically a :term:`dimensionality reduction<Dimensionality Reduction>` technique.
     correction : "bonferroni" | "fdr", default "bonferroni"
         Correction type for multivariate data. Either 'bonferroni' or 'fdr' (False
         Discovery Rate).
@@ -85,7 +85,7 @@ class DriftKS(BaseDriftUnivariate):
     @preprocess_x
     def score(self, x: ArrayLike) -> tuple[NDArray[np.float32], NDArray[np.float32]]:
         """
-        Compute KS scores and statistics per feature.
+        Compute KS scores and :term:Statistics` per feature.
 
         Parameters
         ----------
@@ -95,7 +95,7 @@ class DriftKS(BaseDriftUnivariate):
         Returns
         -------
         tuple[NDArray, NDArray]
-            Feature level p-values and KS statistic
+            Feature level :term:p-values and KS statistic
         """
         x = to_numpy(x)
         x = x.reshape(x.shape[0], -1)

dataeval/_internal/detectors/drift/mmd.py

@@ -24,14 +24,14 @@ from .torch import GaussianRBF, get_device, mmd2_from_kernel_matrix
 @dataclass(frozen=True)
 class DriftMMDOutput(DriftBaseOutput):
     """
-    Output class for :class:`DriftMMD` drift detector
+    Output class for :class:`DriftMMD` :term:`drift<Drift>` detector
 
     Attributes
     ----------
     is_drift : bool
         Drift prediction for the images
     threshold : float
-        P-value used for significance of the permutation test
+        :term:`P-Value` used for significance of the permutation test
     p_val : float
         P-value obtained from the permutation test
     distance : float
@@ -49,14 +49,14 @@ class DriftMMDOutput(DriftBaseOutput):
 
 class DriftMMD(BaseDrift):
     """
-    Maximum Mean Discrepancy (MMD) data drift detector using a permutation test.
+    :term:`Maximum Mean Discrepancy (MMD) Drift Detection` algorithm using a permutation test.
 
     Parameters
     ----------
     x_ref : ArrayLike
         Data used as reference distribution.
     p_val : float | None, default 0.05
-        p-value used for significance of the statistical test for each feature.
+        :term:`P-value` used for significance of the statistical test for each feature.
         If the FDR correction method is used, this corresponds to the acceptable
         q-value.
     x_ref_preprocessed : bool, default False
@@ -69,7 +69,7 @@ class DriftMMD(BaseDrift):
         or via reservoir sampling with ReservoirSamplingUpdateStrategy.
     preprocess_fn : Callable | None, default None
         Function to preprocess the data before computing the data drift metrics.
-        Typically a dimensionality reduction technique.
+        Typically a :term:`dimensionality reduction<Dimensionality Reduction>` technique.
     kernel : Callable, default GaussianRBF
         Kernel used for the MMD computation, defaults to Gaussian RBF kernel.
     sigma : ArrayLike | None, default None
@@ -132,7 +132,7 @@ class DriftMMD(BaseDrift):
     @preprocess_x
     def score(self, x: ArrayLike) -> tuple[float, float, float]:
         """
-        Compute the p-value resulting from a permutation test using the maximum mean
+        Compute the :term:`p-value<P-Value>` resulting from a permutation test using the maximum mean
         discrepancy as a distance measure between the reference data and the data to
         be tested.
 
@@ -145,7 +145,7 @@ class DriftMMD(BaseDrift):
         -------
         tuple(float, float, float)
             p-value obtained from the permutation test, MMD^2 between the reference and test set,
-            and MMD^2 threshold above which drift is flagged
+            and MMD^2 threshold above which :term:`drift<Drift>` is flagged
         """
         x = as_numpy(x)
         x_ref = torch.from_numpy(self.x_ref).to(self.device)
@@ -161,7 +161,7 @@ class DriftMMD(BaseDrift):
         # compute distance threshold
         idx_threshold = int(self.p_val * len(mmd2_permuted))
         distance_threshold = torch.sort(mmd2_permuted, descending=True).values[idx_threshold]
-        return p_val.numpy().item(), mmd2.numpy().item(), distance_threshold.numpy()
+        return p_val.numpy().item(), mmd2.numpy().item(), distance_threshold.numpy().item()
 
     @set_metadata("dataeval.detectors")
     @preprocess_x
@@ -179,7 +179,8 @@ class DriftMMD(BaseDrift):
         Returns
         -------
         DriftMMDOutput
-            Output class containing the drift prediction, p-value, threshold and MMD metric.
+            Output class containing the :term:`drift<Drift>` prediction, :term:`p-value<P-Value>`,
+            threshold and MMD metric.
         """
         # compute drift scores
         p_val, dist, distance_threshold = self.score(x)

dataeval/_internal/detectors/drift/torch.py

@@ -102,7 +102,7 @@ def predict_batch(
     preprocess_fn : Callable | None, default None
         Optional preprocessing function for each batch.
     dtype : np.dtype | torch.dtype, default np.float32
-        Model output type, either a numpy or torch dtype, e.g. np.float32 or torch.float32.
+        Model output type, either a :term:`NumPy` or torch dtype, e.g. np.float32 or torch.float32.
 
     Returns
     -------
@@ -179,7 +179,7 @@ def preprocess_drift(
     batch_size : int, default 1e10
         Batch size used during prediction.
     dtype : np.dtype | torch.dtype, default np.float32
-        Model output type, either a numpy or torch dtype, e.g. np.float32 or torch.float32.
+        Model output type, either a :term:`NumPy` or torch dtype, e.g. np.float32 or torch.float32.
 
     Returns
     -------

dataeval/_internal/detectors/drift/uncertainty.py

@@ -34,7 +34,7 @@ def classifier_uncertainty(
     x : np.ndarray
         Batch of instances.
     model_fn : Callable
-        Function that evaluates a classification model on x in a single call (contains
+        Function that evaluates a :term:`classification<Classification>` model on x in a single call (contains
         batching logic if necessary).
     preds_type : "probs" | "logits", default "probs"
         Type of prediction output by the model. Options are 'probs' (in [0,1]) or
@@ -73,9 +73,9 @@ class DriftUncertainty:
     x_ref : ArrayLike
         Data used as reference distribution.
     model : Callable
-        Classification model outputting class probabilities (or logits)
+        :term:`Classification` model outputting class probabilities (or logits)
     p_val : float, default 0.05
-        p-value used for the significance of the test.
+        :term:`P-Value` used for the significance of the test.
     x_ref_preprocessed : bool, default False
         Whether the given reference data ``x_ref`` has been preprocessed yet.
         If ``True``, only the test data ``x`` will be preprocessed at prediction time.
@@ -145,6 +145,7 @@ class DriftUncertainty:
         Returns
         -------
         DriftUnvariateOutput
-            Dictionary containing the drift prediction, p-value, and threshold statistics.
+            Dictionary containing the drift prediction, :term:`p-value<P-Value>`, and threshold
+            statistics.
         """
         return self._detector.predict(x)

dataeval/_internal/detectors/duplicates.py

@@ -37,7 +37,7 @@ class DuplicatesOutput(Generic[TIndexCollection], OutputMetadata):
 
 class Duplicates:
     """
-    Finds the duplicate images in a dataset using xxhash for exact duplicates
+    Finds the duplicate images in a dataset using xxhash for exact :term:`duplicates<Duplicates>`
     and pchash for near duplicates
 
     Attributes

dataeval/_internal/detectors/ood/ae.py

@@ -24,12 +24,12 @@ from dataeval._internal.output import set_metadata
 
 class OOD_AE(OODBase):
     """
-    Autoencoder based out-of-distribution detector.
+    Autoencoder-based :term:`out of distribution<Out-of-distribution (OOD)>` detector.
 
     Parameters
     ----------
     model : AE
-        An Autoencoder model.
+       An :term:`autoencoder<Autoencoder>` model.
     """
 
     def __init__(self, model: AE) -> None:

dataeval/_internal/detectors/ood/aegmm.py

@@ -30,7 +30,7 @@ class OOD_AEGMM(OODGMMBase):
     Parameters
     ----------
     model : AEGMM
-        An AEGMM model.
+       An AEGMM model.
     """
 
     def __init__(self, model: AEGMM) -> None:
@@ -53,7 +53,7 @@ class OOD_AEGMM(OODGMMBase):
     @set_metadata("dataeval.detectors")
     def score(self, X: ArrayLike, batch_size: int = int(1e10)) -> OODScoreOutput:
         """
-        Compute the out-of-distribution (OOD) score for a given dataset.
+        Compute the :term:`out of distribution<Out-of-distribution (OOD)>` score for a given dataset.
 
         Parameters
         ----------

dataeval/_internal/detectors/ood/base.py

@@ -32,7 +32,7 @@ class OODOutput(OutputMetadata):
     Attributes
     ----------
     is_ood : NDArray
-        Array of images that are detected as out of distribution
+        Array of images that are detected as :term:Out-of-Distribution (OOD)`
     instance_score : NDArray
         Instance score of the evaluated dataset
     feature_score : NDArray | None
@@ -109,7 +109,7 @@ class OODBase(ABC):
     @abstractmethod
     def score(self, X: ArrayLike, batch_size: int = int(1e10)) -> OODScoreOutput:
         """
-        Compute the out-of-distribution (OOD) scores for a given dataset.
+        Compute the :term:`out of distribution<Out-of-distribution (OOD)>` scores for a given dataset.
 
         Parameters
         ----------
@@ -182,7 +182,7 @@ class OODBase(ABC):
         ood_type: Literal["feature", "instance"] = "instance",
     ) -> OODOutput:
         """
-        Predict whether instances are out-of-distribution or not.
+        Predict whether instances are :term:`out of distribution<Out-of-distribution (OOD)>` or not.
 
         Parameters
         ----------

dataeval/_internal/detectors/ood/llr.py

@@ -35,7 +35,7 @@ def build_model(
     Parameters
     ----------
     dist
-        TensorFlow distribution.
+        :term:`TensorFlow` distribution.
     input_shape
         Input shape of the model.
     filepath
@@ -230,7 +230,7 @@ class OOD_LLR(OODBase):
         batch_size: int = int(1e10),
     ) -> NDArray:
         """
-        Compute log probability of a batch of instances under the generative model.
+        Compute log probability of a batch of instances under the :term:`generative model<Generative Model>`.
         """
         logp_fn = partial(dist.log_prob, return_per_feature=return_per_feature)
         # TODO: TBD: can this be any of the other types from predict_batch? i.e. tf.Tensor or tuple
@@ -269,7 +269,7 @@ class OOD_LLR(OODBase):
         return_per_feature
             Return likelihood ratio per feature.
         batch_size
-            Batch size for the generative model evaluations.
+            Batch size for the :term:`generative model<Generative Model>` evaluations.
 
         Returns
         -------

dataeval/_internal/detectors/ood/vae.py

@@ -45,7 +45,7 @@ class OOD_VAE(OODBase):
 
     >>> metric.fit(dataset, threshold_perc=85, batch_size=128, verbose=False)
 
-    Detect out of distribution samples at the 'feature' level
+    Detect :term:`out of distribution<Out-of-Distribution (OOD)>` samples at the 'feature' level
 
     >>> result = metric.predict(dataset, ood_type="feature")
     """

dataeval/_internal/detectors/ood/vaegmm.py

@@ -57,7 +57,7 @@ class OOD_VAEGMM(OODGMMBase):
     @set_metadata("dataeval.detectors")
     def score(self, X: ArrayLike, batch_size: int = int(1e10)) -> OODScoreOutput:
         """
-        Compute the out-of-distribution (OOD) score for a given dataset.
+        Compute the out of distributuion<Out-of-distribution (OOD)>` score for a given dataset.
 
         Parameters
         ----------

dataeval/_internal/detectors/outliers.py

@@ -27,7 +27,7 @@ class OutliersOutput(Generic[TIndexIssueMap], OutputMetadata):
     Attributes
     ----------
     issues : dict[int, dict[str, float]] | list[dict[int, dict[str, float]]]
-        Indices of image outliers with their associated issue type and calculated values.
+        Indices of image Outliers with their associated issue type and calculated values.
 
     - For a single dataset, a dictionary containing the indices of outliers and
       a dictionary showing the issues and calculated values for the given index.
@@ -69,7 +69,7 @@ def _get_outlier_mask(
 
 class Outliers:
     r"""
-    Calculates statistical outliers of a dataset using various statistical tests applied to each image
+    Calculates statistical Outliers of a dataset using various statistical tests applied to each image
 
     Parameters
     ----------
@@ -86,7 +86,7 @@ class Outliers:
 
     See Also
     --------
-    Duplicates
+    :term:`Duplicates`
 
     Note
     ----
@@ -162,7 +162,7 @@ class Outliers:
         self, stats: OutlierStatsOutput | DatasetStatsOutput | Sequence[OutlierStatsOutput]
     ) -> OutliersOutput:
         """
-        Returns indices of outliers with the issues identified for each
+        Returns indices of Outliers with the issues identified for each
 
         Parameters
         ----------
@@ -238,7 +238,7 @@ class Outliers:
     )
     def evaluate(self, data: Iterable[ArrayLike]) -> OutliersOutput[IndexIssueMap]:
         """
-        Returns indices of outliers with the issues identified for each
+        Returns indices of Outliers with the issues identified for each
 
         Parameters
         ----------

dataeval/_internal/metrics/balance.py

@@ -15,12 +15,12 @@ from dataeval._internal.output import OutputMetadata, set_metadata
 @dataclass(frozen=True)
 class BalanceOutput(OutputMetadata):
     """
-    Output class for :func:`balance` bias metric
+    Output class for :func:`balance` :term:`Bias` metric
 
     Attributes
     ----------
     balance : NDArray[np.float64]
-        Estimate of mutual information between metadata factors and class label
+        Estimate of :term:`mutual information<Mutual Information (MI)>` between metadata factors and class label
     factors : NDArray[np.float64]
         Estimate of inter/intra-factor mutual information
     classwise : NDArray[np.float64]
@@ -55,7 +55,7 @@ def validate_num_neighbors(num_neighbors: int) -> int:
 @set_metadata("dataeval.metrics")
 def balance(class_labels: ArrayLike, metadata: Mapping[str, ArrayLike], num_neighbors: int = 5) -> BalanceOutput:
     """
-    Mutual information (MI) between factors (class label, metadata, label/image properties)
+    :term:`Mutual information (MI)` between factors (class label, metadata, label/image properties)
 
     Parameters
     ----------
@@ -70,7 +70,7 @@ def balance(class_labels: ArrayLike, metadata: Mapping[str, ArrayLike], num_neig
     Returns
     -------
     BalanceOutput
-        (num_factors+1) x (num_factors+1) estimate of mutual information
+        (num_factors+1) x (num_factors+1) estimate of :term:`mutual information<Mutual Information (MI)>`
         between num_factors metadata factors and class label. Symmetry is enforced.
 
     Note
@@ -83,7 +83,7 @@ def balance(class_labels: ArrayLike, metadata: Mapping[str, ArrayLike], num_neig
 
     Example
     -------
-    Return balance (mutual information) of factors with class_labels
+    Return :term:`balance<Balance>` (:term:`mutual information<Mutual Information (MI)>`) of factors with class_labels
 
     >>> bal = balance(class_labels, metadata)
     >>> bal.balance

dataeval/_internal/metrics/ber.py

@@ -1,7 +1,7 @@
 """
 This module contains the implementation of the
 FR Test Statistic based estimate and the
-KNN based estimate for the Bayes Error Rate
+KNN based estimate for the :term:`Bayes error rate<Bayes Error Rate (BER)>`
 
 Learning to Bound the Multi-class Bayes Error (Th. 3 and Th. 4)
 https://arxiv.org/abs/1811.06419
@@ -30,7 +30,7 @@ class BEROutput(OutputMetadata):
     Attributes
     ----------
     ber : float
-        The upper bounds of the Bayes Error Rate
+        The upper bounds of the :term:`Bayes error rate<Bayes Error Rate (BER)>`
     ber_lower : float
         The lower bounds of the Bayes Error Rate
     """
@@ -40,7 +40,7 @@ class BEROutput(OutputMetadata):
 
 
 def ber_mst(X: NDArray, y: NDArray) -> tuple[float, float]:
-    """Calculates the Bayes Error Rate using a minimum spanning tree
+    """Calculates the :term:`Bayes error rate<Bayes Error Rate (BER)>` using a minimum spanning tree
 
     Parameters
     ----------
@@ -65,7 +65,7 @@ def ber_mst(X: NDArray, y: NDArray) -> tuple[float, float]:
 
 
 def ber_knn(X: NDArray, y: NDArray, k: int) -> tuple[float, float]:
-    """Calculates the Bayes Error Rate using K-nearest neighbors
+    """Calculates the :term:`Bayes error rate<Bayes Error Rate (BER)>` using K-nearest neighbors
 
     Parameters
     ----------
@@ -114,12 +114,12 @@ BER_FN_MAP = {"KNN": ber_knn, "MST": ber_mst}
 @set_metadata("dataeval.metrics")
 def ber(images: ArrayLike, labels: ArrayLike, k: int = 1, method: Literal["KNN", "MST"] = "KNN") -> BEROutput:
     """
-    An estimator for Multi-class Bayes Error Rate using FR or KNN test statistic basis
+    An estimator for Multi-class :term:`Bayes error rate<Bayes Error Rate (BER)>` using FR or KNN test statistic basis
 
     Parameters
     ----------
     images : ArrayLike (N, ... )
-        Array of images or image embeddings
+        Array of images or image :term:`embeddings<Embeddings>`
     labels : ArrayLike (N, 1)
         Array of labels for each image or image embedding
     k : int, default 1