Closed wlach closed 5 years ago
This might be a fun project for a motivated contributor to work on.
Is this issue still open??
@harshagarwal7272 it is, but I'm less certain it's a good idea now after getting more familiar with the error aggregates data. I don't think it quite fits the pattern described in their paper. Still not opposed to people experimenting in this area, but I think other things might be better to work on.
Let's close this out for now.
Stanford has some techniques for smoothing noisy data to generate a signal, looks like the techniques might be useful for mission control:
http://futuredata.stanford.edu/asap/#code
We might be able to just plug their library in, though I'd like to spend some time to understand what it's doing first. Is this something we want in addition to the percentile filtering which was just added?