The shapes of spectral profiles contain crucial information about the structure of the solar atmosphere which generates them. However, the relationship between profiles and atmospheric structure is not simple; rather different atmospheres can produce very similar spectral profiles. Thus, in order to properly interpret observations, one needs to have a solid understanding of how atmospheric parameters like temperature, velocities, and magnetic fields impact the profiles' formation.
Forward modeling, i.e. computing and studying synthetic spectra from realistic atmospheric models, is a crucial tool in this endeavour, allowing a direct comparison between spectral profiles and the structure of the atmosphere which they arise from. A modern rMHD simulation of the solar atmosphere might have 512x512x512 or more grid cells, giving rise to several hundred thousand spectral profiles for each time-step. Such a number of profiles invites the application of clustering techniques to group similar spectra together, and in turn reveal the commonalities and differences in the atmospheric conditions of their formation.
In this talk I will give an introduction to two specific clustering techniques, k-means and k-Shape, and discuss their application to synthetic Stokes profiles from a Bifrost snapshot.
This Friday colloquium will be hybrid. Attendees can therefore participate either in-person or via Zoom. Please join via Zoom at
https://uio.zoom.us/j/69001043754?pwd=cEJpbVE5ci9PdWNtRld2TDNNcGtKdz09
Meeting ID: 690 0104 3754
Passcode: PeiseStua3
Attendees will be muted during the colloquium, but will have the opportunity to ask questions at the end by clicking on the "raise hand” button (or send a request via chat).