2

I have some data points and I fitted a function (2nd order polynomial here) to the data. The algorithm (scipy.optimize.curve_fit) gave me the optimal parameters and the covariance matrix of those parameters. In the documentation, it says that the square roots of the diagonal entries give the errors for the parameters.

Now I that I have the function $f$ fitted to points $0 < x$, I would like to find the value of $f(0)$ and its error. $f(0)$ is easy, but how do I acquire an estimate for the error?

Martin Ueding
  • 141
  • 7
  • The answer at http://stats.stackexchange.com/questions/9131/obtaining-a-formula-for-prediction-limits-in-a-linear-model gives a formula. – whuber Jul 01 '14 at 18:27

1 Answers1

0

I ended up using bootstrap, that gave me a decent looking error estimate.

Martin Ueding
  • 141
  • 7