Each least squares fit provides a covariance matrix for the identified parameters. I do the simple thing, and take the diagonal of that to get an uncertainty for each parameter and then run with that, including computing the mean gains across multiple trials. The mean calculation may need to take into account the covariance instead of just the variance.
moorepants
commented
9 years ago
Each least squares fit provides a covariance matrix for the identified parameters. I do the simple thing, and take the diagonal of that to get an uncertainty for each parameter and then run with that, including computing the mean gains across multiple trials. The mean calculation may need to take into account the covariance instead of just the variance.