Koposov, Lang, Rix, and I debated methods for finding streams or substructure in high-dimensional stellar data sets, with Koposov defending brute-force foreground/background modeling, and Lang defending density estimation as a search heuristic. We all agreed that the best way to confirm a stream is with foreground/background modeling, so we were sympathetic with Koposov, but there are issues with speed, as there are in all brute-force methods.
No comments:
Post a Comment