Least-Squares Straight-Line Fitting Algorithm with Automatic Error Determination

摘要

We demonstrate a new generalized least squares fitting method which can be used to estimate the slope of the best fitting straight line that results when two separate data sets which are expected to be linearly correlated are subject to different uncertainties in their measurements. The algorithm determines not only the optimum slope, but also produces estimates of the intrinsic errors associated with each data set. It requires almost no initial information about the errors in each technique.