help fine-tuning parameters

[Sam Mckenzie]

Hello, 

I have been flailing around parameter space trying to find some combination that is tolerable.

My problem is that the template matching is too permissive and accepts MUA. It is rare that I can actually have two units merged into one, the issue seems to be a real unit accepting lots of noise. I know the signal is recoverable, because if I run klustakwik on my noise clusters, I can recover perfect units (see attached). 

Merging is easy, so i would rather over cluster. How can I make more templates each with more restrictive matching thresholds. My hunch is that the issue has to do with the algorithm de-emphasizing the real voltage value of the spike. 

thank you

sam

[Sam Mckenzie]

and another that circus considered a single unit but klustakwik resolved.  I am considering running these two algorithms in series as a rule (klustakwik does not scale well), but I would love to have circus do the job as a stand alone

--
You received this message because you are subscribed to a topic in the Google Groups "SpyKING CIRCUS" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/spyking-circus-users/auOL_7BCE9E/unsubscribe.
To unsubscribe from this group and all its topics, send an email to spyking-circus-u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/spyking-circus-users/b30f2b02-c12e-4804-b64a-85bf700baa85n%40googlegroups.com.