These notes discuss, in a style intended for physicists, how to average data and fit it to some functional form. I try to make clear what is being calculated, what assumptions are being made, and to give a derivation of results rather than just quote them. The aim is put a lot useful pedagogical material together in a convenient place. This manuscript is a substantial enlargement of lecture notes I prepared for the Bad Honnef School on "Efficient Algorithms in Computational Physics", September 10-14, 2012.
- Pub Date:
- October 2012
- Physics - Data Analysis;
- Statistics and Probability;
- Condensed Matter - Statistical Mechanics;
- Physics - Computational Physics
- 78 pages, 15 figures. Now includes a discussion of overfitting, including both Maximum Likelihood and Bayesian approaches