Closed khider closed 1 year ago
Nice didactic notebook on clustering in general, and for timeseries in particular. It still sounds to me like a hammer to swat a fly, because at the end of the day you end up choosing the parameter values that confirm the "eye test". That is a lot of CPU time and lines of code to simply confirm your bias! However, all of data analysis is more art than science, so perhaps this is no different.
I remember you saying that .outliers()
can be used to flag regime transitions, and that might be a more interesting application for paleo folks. Is there a quick example you could sneak in there, or would you rather keep that a PaleoBook?
It can. But probably a better Books than tutorials.
Creates a notebook tutorial for outlier detection in Pyleoclim. Walks through the general principles of clustering (VERY IMPORTANT TO UNDERSTAND WHAT PYLEOCLIM DOES UNDER THE HOOD - DO NOT SKIP) and shows an example in Pyleoclim.