Default value for max_distance searching for matches #931
Replies: 1 comment
-
@MiguelGarcaoSilva Thank you for your question and welcome to the STUMPY community!
To be honest, this is just a super crude/naive way to identify motif distances that are (possibly) exceptional compared with all other distances. We (naively) assume that distribution of matrix profile distances follow a normal distribution and that all distances below 2 stddevs from the mean are "interesting"/"significant" to look at first. Since the time series can vary drastically from use case to use case, this naive default value adjusts according to the time series data being used. Also, choosing the Note that this is a sane/reasonable starting point but you are free to specify your own |
Beta Was this translation helpful? Give feedback.
-
Hello,
I've noticed the default value for the max_distance parameter in the stumpy.motifs function is set to np.nanmax([np.nanmean(D) - 2.0 * np.nanstd(D), np.nanmin(D)]) . Is there a specific reason to use this value as default? Are there any published papers recommending it?
Thanks in advance!
Beta Was this translation helpful? Give feedback.
All reactions