dataeval.detectors.linters.Outliers

class dataeval.detectors.linters.Outliers(use_dimension=True, use_pixel=True, use_visual=True, outlier_method='modzscore', outlier_threshold=None)

Calculates statistical outliers of a dataset using various statistical tests applied to each image.

Parameters:
use_dimension : bool, default True

If True, use dimension statistics to identify outliers

use_pixel : bool, default True

If True, use pixel statistics to identify outliers

use_visual : bool, default True

If True, use visual statistics to identify outliers

outlier_method : ["modzscore" | "zscore" | "iqr"], optional - default "modzscore"

Statistical method used to identify outliers

outlier_threshold : float, optional - default None

Threshold value for the given outlier_method, above which data is considered an outlier - uses method specific default if None

stats

Various stats output classes that hold the value of each metric for each image

Type:

tuple[DimensionStatsOutput, PixelStatsOutput, VisualStatsOutput]

See also

Duplicates

Note

There are 3 different statistical methods:

  • zscore

  • modzscore

  • iqr

The z score method is based on the difference between the data point and the mean of the data. The default threshold value for zscore is 3.
Z score = \(|x_i - \mu| / \sigma\)
The modified z score method is based on the difference between the data point and the median of the data. The default threshold value for modzscore is 3.5.
Modified z score = \(0.6745 * |x_i - x̃| / MAD\), where \(MAD\) is the median absolute deviation
The interquartile range method is based on the difference between the data point and the difference between the 75th and 25th qartile. The default threshold value for iqr is 1.5.
Interquartile range = \(threshold * (Q_3 - Q_1)\)

Examples

Initialize the Outliers class:

>>> outliers = Outliers()

Specifying an outlier method:

>>> outliers = Outliers(outlier_method="iqr")

Specifying an outlier method and threshold:

>>> outliers = Outliers(outlier_method="zscore", outlier_threshold=3.5)
evaluate(data)

Returns indices of Outliers with the issues identified for each

Parameters:
data : Iterable[Array], shape - (C, H, W)

A dataset of images in an Array format

Returns:

Output class containing the indices of outliers and a dictionary showing the issues and calculated values for the given index.

Return type:

OutliersOutput

Example

Evaluate the dataset:

>>> outliers = Outliers(outlier_method="zscore", outlier_threshold=3.5)
>>> results = outliers.evaluate(outlier_images)
>>> list(results.issues)
[10, 12]
>>> results.issues[10]
{'contrast': 1.2499999999203126, 'entropy': 0.21278774841317422, 'zeros': 0.054931640625}
from_stats(stats: OutlierStatsOutput) → dataeval.outputs.OutliersOutput[dataeval.outputs._linters.IndexIssueMap]
from_stats(stats: collections.abc.Sequence[OutlierStatsOutput]) → dataeval.outputs.OutliersOutput[list[dataeval.outputs._linters.IndexIssueMap]]

Returns indices of Outliers with the issues identified for each.

Parameters:
stats : OutlierStatsOutput | ImageStatsOutput | Sequence[OutlierStatsOutput]

The output(s) from a dimensionstats, pixelstats, or visualstats metric analysis or an aggregate ImageStatsOutput

Returns:

Output class containing the indices of outliers and a dictionary showing the issues and calculated values for the given index.

Return type:

OutliersOutput

See also

dimensionstats, pixelstats, visualstats

Example

Evaluate the dataset:

>>> outliers = Outliers(outlier_method="zscore", outlier_threshold=3.5)
>>> results = outliers.from_stats([stats1, stats2])
>>> len(results)
2
>>> results.issues[0]
{10: {'entropy': 0.2128, 'zeros': 0.05493}, 12: {'entropy': 0.2128, 'std': 0.00536, 'var': 2.87e-05, 'zeros': 0.05493}}
>>> results.issues[1]
{}