VACILT/student_project

understanding the regression

Closed this issue · 2 comments

stoeckel4code commented

  1. I don't get how the coefficients are are depending on more than one variable.
    I thought the regression must be y(a,b)=z0+z1x1(a,b)+z2x2(a,b)+ ... + zn*xn(a,b), where zn are the coefficients, which only have one dimension.

  2. How can the unit be "%"?

kuchaale commented
  1. Regression coefficients are 2D because regressed time-series are 2D+time and regressors are 1D, i.e. y(x,y,t) = z0(x,y) + z1(x,y)*x1(t) + ...
  2. Because the regressed time-series are expressed in % in terms of climatology (see here).
stoeckel4code commented

  1. So the regression is done for every point of (x,y) and the regression is only over the time? Am I right?

  2. ok, makes sence.

  • 👍1