This is an unofficial python implementation of DRAG algorithm, which efficiently searches for time series discords. DRAG was first introduced in [1], but was not named, hence the name in [2].
from drag import DRAG
window_size = 100
discord_defining_range = 10.
C, C_dist = DRAG(X, window_size, discord_defining_range)- Yankov, Dragomir, Eamonn Keogh, and Umaa Rebbapragada. "Disk aware discord discovery: Finding unusual time series in terabyte sized datasets." Knowledge and Information Systems 17 (2008): 241-262.
- Nakamura, Takaaki, et al. "Merlin: Parameter-free discovery of arbitrary length anomalies in massive time series archives." 2020 IEEE international conference on data mining (ICDM). IEEE, 2020.