So I'm working with some data and I've noticed that the best fitting trendline is a polynomial function rather than a linear equation. I've found a tutorial that's a couple years old that allowed me to at least generate the LINEST array for the regression and the TINV values for certain percentiles, I don't know how to plug them into the bounding formulas I found on this site, since they were for linear equations. So for an example, here is one of the arrays I've generated for it so far:
So I'm working with some data and I've noticed that the best fitting trendline is a polynomial function rather than a linear equation. I've found a tutorial that's a couple years old that allowed me to at least generate the LINEST array for the regression and the TINV values for certain percentiles, I don't know how to plug them into the bounding formulas I found on this site, since they were for linear equations. So for an example, here is one of the arrays I've generated for it so far:
-3.27055E-14 1.77069E-10 -3.43246E-7 0.00028586 -0.108005 58.81177
5.04074E-14 2.34805E-10 3.65532E-7 0.000217063 0.03924 1.122581
0.336886741 8.287261549 #N/A #N/A #N/A #N/A
11.78648184 116 #N/A #N/A #N/A #N/A
4047.401486 7966.729661 #N/A #N/A #N/A #N/A
So I know that the second row are the standard errors of the coefficients, but how do I go about incorporating them into the formula? The formula I found allows for only one slope standard error and one intercept standard error, but I don't know how to go about adding in the others. Any and all help would be appreciated.
So I know that the second row are the standard errors of the coefficients, but how do I go about incorporating them into the formula? The formula I found allows for only one slope standard error and one intercept standard error, but I don't know how to go about adding in the others. Any and all help would be appreciated.