Okay, the heading is probably confusing, but I am looking for the name of an algorithm (if any exists) or what can I use that will detect patterns in continuous data. The example image should explain better (sadly I can't embed it in the post). You can think of the x axis as time or any ordered indexes, and the y axis are the values I care about. They (the y-values) will mostly be with a very small variance, with occasional peaks. What I want to detect are these peaks (encircled in red in the image, or more specifically, I only need the start and end index of the given peak - I've marked this with green squares)
A much simpler way to think about it is with the data simply a list e.g.
[8, 10, 8, 12, 9, 12, 7, 150, 160, 155, 150, 160, 170, 9, 8, 12, 10, ...]
and what I care about are the indexes of the element values in the 150-170s range.
Now this seems pretty simple and I can probably write a program to detect these myself, but I though there must be function(s)/packages that can achieve this. I simply have no idea what to look for...
So yeah any directions on what to look for, or what can be used to achieve this will be helpful. I'm using Python if it matters.
Aucun commentaire:
Enregistrer un commentaire