Skip to content

Outliers

Functions for outlier detection

get_outlier_segments(is_outlier, round_to='1s')

List time segments of outliers

Parameters:

Name Type Description Default
is_outlier Series

Time series of samples marked as outlier.

 required

Returns:

Type Description
list[tuple[Timestamp, Timestamp]]

List of time ranges of outliers.

guess_mean_and_std(g, interval='1min')

Guess mean and standard deviation of signal with outliers

Parameters:

Name Type Description Default
g Series

Input graviy time series.

 required

Returns:

Name Type Description
g_mean float

Estimated mean.
g_std float

Standard deviation.

join_close_segments(is_outlier, min_gap)

Join outlier segments that are close together

Parameters:

Name Type Description Default
is_outlier Series

Time series of samples marked as outlier.

 required
min_gap str

Minimum time gap between outliers to allow. Outliers that are closer together are joined into one outlier segment.

 required

Returns:

Name Type Description
is_outlier Series

Time series of samples marked as outlier.

reject_by_sigma_threshold(g, sigma_threshold, g_true=None, g_std=None)

Reject samples by a sigma criterion

Parameters:

Name Type Description Default
g Series

Input graviy time series.

 required
sigma_threshold float

Maximum allowed standard deviations.

 required
g_true float

Reference value.

 None
g_std float

Reference standard deviation.

 None

Returns:

Name Type Description
is_outlier Series

Time series of samples marked as outliers.

reject_by_threshold(g, threshold, g_true=None)

Reject samples that deviate by more than a threshold value

Parameters:

Name Type Description Default
g Series

Input gravity time series.

 required
threshold float

Maximum allowed deviation from reference value.

 required
g_true float

Reference value to calculate deviation from. If unspecified, it is estimated from the median of the first 1000 samples.

 None

Returns:

Name Type Description
is_outlier Series

Time series of samples marked as outliers.

reject_neighbors(is_outlier, num)

Reject nearest neighbors of outliers

Parameters:

Name Type Description Default
is_outlier Series

Series of samples marked as outlier.

 required
num int

Number of neighbors on either side to reject.

 required

Returns:

Name Type Description
is_outlier Series

Time series of samples marked as outlier with additionl neighbors marked.