search

bjartelund / Thermoslope

GNU Affero General Public License v3.0
1 stars 0 forks source link

Std. errors should be kept between the steps #2

Closed bjartelund closed 3 years ago

bjartelund commented 4 years ago

The concept is three consecutive regressions, and from each set measurements there will be errors. It would be better if the standard errors were used in the regressions to give weights to the data points, so that poorly defined curves do not influence too much.

bjartelund commented 3 years ago

PR #11 Partially solves this by weighting the arrhenius plot by the covariance from the fitting.