Uploaded image for project: 'Data Management'
  1. Data Management
  2. DM-19292

Compare detection rates to Gaussian noise for HiTS processing

    Details

    • Story Points:
      8
    • Epic Link:
    • Team:
      Alert Production

      Description

      This ticket is to effectively repeat the relevant portions of the DMTN-006/DMTN-021 analysis for our current HiTS processing to bound our false positive rates. This ticket is simply to perform the work in a notebook, but it is expected that future work will incorporate relevant portions into the metrics system.

        Attachments

          Issue Links

            Activity

            Hide
            ebellm Eric Bellm added a comment -

            Hi Gabor Kovacs, can you take a look at this?

            Show
            ebellm Eric Bellm added a comment - Hi Gabor Kovacs , can you take a look at this?
            Hide
            swinbank John Swinbank added a comment -

            Hey Gabor Kovacs, Eric Bellm — looks like this one has been in review for a month now, with no obvious progress. Any updates? It would be great to close it out before the end of May.

            Show
            swinbank John Swinbank added a comment - Hey Gabor Kovacs , Eric Bellm — looks like this one has been in review for a month now, with no obvious progress. Any updates? It would be great to close it out before the end of May.
            Hide
            gkovacs Gabor Kovacs added a comment -

            Minor comments:

            I think, in cell 12: y axis is (total) number of diasources not ccdvisits.

            I'd recommend adding 1-2 extra sentences for more context and referring to additional docs beside DMTN-021 if available:

            • We have two issues here: that DMTN-021 numbers are not reproduced (50 vs. 180) and that both numbers are much higher than what the noise_detections formula predicts (1.5).
            • Does the 5sigma detection exactly mean the same as the noise_detections formula threshold==5 ? Is there any docs about the source detection ?
            • How is fluxerr derived? When should we expect SNR == detection sigma ? Is there any docs about the flux measurement (errors) ?
            Show
            gkovacs Gabor Kovacs added a comment - Minor comments: I think, in cell 12: y axis is (total) number of diasources not ccdvisits. I'd recommend adding 1-2 extra sentences for more context and referring to additional docs beside DMTN-021 if available: We have two issues here: that DMTN-021 numbers are not reproduced (50 vs. 180) and that both numbers are much higher than what the noise_detections formula predicts (1.5). Does the 5sigma detection exactly mean the same as the noise_detections formula threshold==5 ? Is there any docs about the source detection ? How is fluxerr derived? When should we expect SNR == detection sigma ? Is there any docs about the flux measurement (errors) ?
            Hide
            ebellm Eric Bellm added a comment -

            Thanks Gabor Kovacs. I fixed the axis labels and added a note that we don't expect noise_detection values since there are real astrophysical variables. I deferred the discussion of detection vs. measurement SNR since we don't yet propagate out the relevant peak values from the detection footprints. That will become more important once we get rid of some of the gross failures.

            Show
            ebellm Eric Bellm added a comment - Thanks Gabor Kovacs . I fixed the axis labels and added a note that we don't expect noise_detection values since there are real astrophysical variables. I deferred the discussion of detection vs. measurement SNR since we don't yet propagate out the relevant peak values from the detection footprints. That will become more important once we get rid of some of the gross failures.

              People

              • Assignee:
                ebellm Eric Bellm
                Reporter:
                ebellm Eric Bellm
                Reviewers:
                Gabor Kovacs
                Watchers:
                Eric Bellm, Gabor Kovacs, John Swinbank
              • Votes:
                0 Vote for this issue
                Watchers:
                3 Start watching this issue

                Dates

                • Created:
                  Updated:
                  Resolved: