"Janani Chander" <email@example.com> wrote in message news:firstname.lastname@example.org... > Hi, > > I would like to if there is anyway to increase the number of significant > digits in curve fitting. I tried to fit and got a fittiong co-efficient > as below. > inear model Poly1: > f(x) = p1*x + p2 > where x is normalized by mean 0.0002051 and std 0.0002143 > Coefficients (with 95% confidence bounds): > p1 = 8.119e-05 (8.043e-05, 8.195e-05) > p2 = -43.82 (-43.82, -43.82) > > But for the confidence bound is the same as co efficient, is there anyway > i can increase the significant digits in the parameters?
Note that this is simply _displaying_ the coefficient using a small number of decimal places. The actual coefficients are stored in the fit object in double precision; save/export the fit object to the workspace and use the COEFFVALUES function to retrieve them. The CONFINT function will give you the confidence bounds. You may also need to change the display FORMAT.