Let's say you have a "form" of the estimating equation that you want to use. In your case y = a Ln(x) +c and the problem is to determine values for a and c. In your case Excel has told you that a = 29.745 and c = -101.84.

You could calculate these values yourself using matrices.

We have the matrix equation:

Y = X*A where:

Y is a column matrix containing all of your coverage values

A is a column matrix with just two elements being a and c

X contains two columns one full of 1's and the other containing Ln of your TFNs

i.e.

Now we need to solve this for the matrix A so that we get a and c.

Y = XA

so

so

is what we call a pseudo inverse of X.

I am going to guess that curve fitting in Excel may be the easier option for you?