src.dackar.anomalies.t_score

Created on July, 2025

@author: wangc, mandd

Attributes

logger

Functions

getL(loc, window, array, type)	Method designed to extract a portion of a time series
omega(window, array, Nsamples)	Method designed to extract set of portions of a time series
choice(array, Nsamples)	Method designed to randomly choose Nsamples out of an array (replace is set to True)
t_score(array_front, array_rear)	Method designed to calculate the statistical difference between two arrays using t-test
MMD_test(test_array, omegaSet, iterations, alphaTest, ...)	Method designed to calculate the statistical difference between an array and a set of arrays
event2TStest(E_loc, TS, iterations, alphaTest, ...[, ...])	Method designed to assess temporal correlation of an event E and a time series TS

Module Contents

src.dackar.anomalies.t_score.logger

src.dackar.anomalies.t_score.getL(loc, window, array, type)

Method designed to extract a portion of a time series

Parameters:

• loc – int, reference location in the time series for the selected portion

• window – np.array, size of the selected portion of the time series

• array – np.array, original time series

• type – string, type of the selected portion of the time series (rear or front)

Returns:

np array, size of the selected portion of the time series

Return type:

timeseriesPortion

src.dackar.anomalies.t_score.omega(window, array, Nsamples)

Method designed to extract set of portions of a time series

Parameters:

• window – np.array, size of the selected portion of the time series

• array – np.array, original time series

• Nsamples – int, number of portions to be selected

Returns:

np.array, set of Nsamples portions of a time series

Return type:

omegaSet

src.dackar.anomalies.t_score.choice(array, Nsamples)

Method designed to randomly choose Nsamples out of an array (replace is set to True)

Parameters:

• array – np.array, array of values to be sampled

• Nsamples – int, number of elements of array to be selected

Returns:

np.array, set of Nsamples elements randomly chosen from array

Return type:

sampled

src.dackar.anomalies.t_score.t_score(array_front, array_rear)

Method designed to calculate the statistical difference between two arrays using t-test

Parameters:

• array_front – np.array, first array

• array_rear – np.array, second array

Returns:

float, outcome of t-test

Return type:

tscore

src.dackar.anomalies.t_score.MMD_test(test_array, omegaSet, iterations, alphaTest, alphaOmegaset, printFlag=True)

Method designed to calculate the statistical difference between an array and a set of arrays using the Maximum Mean Discrepancy testing method

Parameters:

• test_array – np.array, array of values to be tested against omegaSet

• omegaSet – np.ndarray or list, population of arrays

• iterations – int, number of iterartions required to compute MMD^2_u

• alphaTest – float, acceptance value for hypothesis testing (single array testing)

• alphaOmegaset – float, acceptance value for hypothesis testing (omegaSet testing)

• printFlag – bool, flag to plot MMD distribution

Returns:

float, p-value obtained from MMD testing testOutput: bool, logical outcome of MMD testing

Return type:

p_value

src.dackar.anomalies.t_score.event2TStest(E_loc, TS, iterations, alphaTest, alphaOmegaset, windowSize, omegaSize, returnMinPval=False)

Method designed to assess temporal correlation of an event E and a time series TS

Parameters:

• E_loc – int, temporal location of event E

• TS – np.array, univariate time series

• iterations – int, number of iterartions required to compute MMD^2_u

• alphaTest – float, acceptance value for hypothesis testing (single array testing)

• alphaOmegaset – float, acceptance value for hypothesis testing (omegaSet testing)

• windowSize – int, size of time window prior and after event occurence

• omegaSize – int, number of samples from time series TS

• returnMinPval – bool, flag that indictae whether to return p-value of MMD assessment

Returns:

string, outcome of the event to timeseries temporal analysis minPval: float, p-value of MMD assessment

Return type:

relation