stumpy.motifs question regarding "cutoff" and "max_distance" parameter? #548
-
Hi, I was hoping someone could help me better understand the "cutoff" parameter in "stumpy.motifs". From docs:
I do understand that max_distance is the maximum distance that a motif can have to it's neighbor. But I don't understand how cutoff is used. Thanks for any help. |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment 6 replies
-
@jmrichardson Thank you for your question and welcome to the STUMPY community.
So, after you've computed the matrix profile, you end up with an array of distances (to their one-nearest-neighbor). Now, say you are interested in finding a motif AND a list of their top matches but, first, you need to scan along your matrix profile and pick a motif. How do you decide which subsequence is a "good" candidate? Typically, you'd choose the subsequence with the smallest matrix profile value as your top motif and then look for "matches" to this motif in your original time series. But after this first subsequence, how would you find your next "candidate motif"? Well, you'd repeat the same process and grab the next smallest matrix profile value and so on. However, let's say the first candidate motif had a matrix profile value of Does that make sense? |
Beta Was this translation helpful? Give feedback.
@jmrichardson Thank you for your question and welcome to the STUMPY community.
So, after you've computed the matrix profile, you end up with an array of distances (to their one-nearest-neighbor). Now, say you are interested in finding a motif AND a list of their top matches but, first, you need to scan along your matrix profile and pick a motif. How do you decide which subsequence is a "good" candidate? Typically, you'd choose the subsequence with the smallest matrix profile value as your top motif and then look for "matches" t…