Homework #6

Due November 10, 2009

This is a follow-on to homework 5. Download the same data file and create the same function to overlay on the data.

You can model each data set as m=i*(1.-exp(-(dl-L0)*(dl-L0)/2./w/w))+b

Here, i is the intensity of the continuum.  b is the detector background. L0 is the center of the line. w is the width of the line.

Create an S Statistic, which is the sum of (di-xi)^2/xi as explained in class. Minimize S by going through each of the parameters one-by-one. For each parameter, follow S until it hits a minimum. Set the parameter to that value, then follow another parameter to its minimum, and so on. Cycle through the parameters until you just don't find much improvement. These are your best fit values. What are they?

Follow each parameter out along its axis until S increases by 1. Those are your error bars. What are they?