Is there any syntax for covariance which uses the unbiased formula (yi-ymean)*(yj-ymean) /n...Monte Carlo method
On Sat, Oct 24, 2020, 21:11 Meghna Karkera <mkarker...@gmail.com> wrote: > Dear Sir > > Now I figured out why there was a difference when I manually computed > covariance and when I use the np. cov syntax in python. > > The formula I used was ((yi-ymean)(yj-ymean))/n...unbiasdness. > > > The formula np.cov uses is ((yi-ymean)(yj-ymean)) /n-1. > > Thanks > > Sincerely > Meghna Raviraj Karkera > > On Mon, Oct 19, 2020, 21:30 Markos <mar...@c2o.pro.br> wrote: > >> Hi Meghna, >> >> I organized some tutorials about Machine Learning for Chemistry with >> Python. >> >> Where you can find some info about the use of Covariance Matrix that may >> help you. >> >> The originals are in Portugueese, but you can read using Google >> Translator: >> >> https://tinyurl.com/yybazx9n >> >> https://tinyurl.com/y29mule6 >> >> https://tinyurl.com/yxtg4tce >> >> Best Regards, >> Markos >> >> Em 19-10-2020 02:23, Meghna Karkera escreveu: >> > Dear Sir >> > >> > I am unable to find the *formula for covariance* used in np.cov syntax >> in >> > PYTHON given in link >> > https://numpy.org/doc/stable/reference/generated/numpy.cov.html. >> Could you >> > please help me out. >> > >> > Thanks >> > Meghna >> > >> > On Tue, Oct 13, 2020 at 11:46 AM Christian Gollwitzer <aurio...@gmx.de> >> > wrote: >> > >> >> Am 13.10.20 um 06:52 schrieb Meghna Karkera: >> >>> Could you let me know what is the back end calculation of this >> covariance >> >>> matrix syntax np.cov >> >>> >> >> You can look it up yourself: Go to the docs >> >> https://numpy.org/doc/stable/reference/generated/numpy.cov.html >> >> >> >> At the right hand side, just right of the function signature, there is >> a >> >> link [source]. Click there and it takes you to the implementation. >> >> >> >> Apparently it is done in straightforward Python. >> >> >> >> Christian >> >> >> >> PS: Snipped a lot of unrelated citation at the bottom >> >> >> >>> On Tue, Oct 13, 2020, 10:14 Bruno P. Kinoshita < >> >> brunodepau...@yahoo.com.br> >> >>> wrote: >> >>> [...] >> >>>> I think the np.cov is from the numpy module (imported/aliased as >> np?). >> >> -- >> >> https://mail.python.org/mailman/listinfo/python-list >> >> >> >> -- >> https://mail.python.org/mailman/listinfo/python-list >> > -- https://mail.python.org/mailman/listinfo/python-list
