What do codes do in Diffpy_CMI?

Skip to first unread message

ali dordaee

Aug 1, 2021, 7:30:16 PMAug 1
to diffpy-users
Dear all, 
 Sorry if my questions look rudimentary. I have three questions:
 1. What is Scipy's LM optimizer matematically and where can I find the mathematical equations for it, if you know a good book I'd really appreciate it. Is it different than Scipy's least-squares method?

 2. The source codes in Github for Diffpy-CMI seem a little vague for me. What I mean is they definitely have used some mathematical equations, that's why they are working fine. I think those equations and algorithms can not be found in the GitHub page. To be more specific, I need to know how EXACTLY and MATHEMATICALLY diffpy-cmi calculates and obtains different parameters(like lattice parameter, particle size, ADP's, etc.). All I know is it definitely uses an iterative method, but if there is an in-situ example or an example which lets us follow through the process step by step(like an slow-motion movie), I'd really appreciate it if you could please introduce it to me. All I need is to know the iterative processes that are going on here.
 3. There are some functions in diffpy-cmi, like Fitrecipe() or PDFcontribution(), Profile(), PDFParser(), etc. which are still a little bit unclear for me. Can I find what exactly they are in somewhere? When I want to use "addVar" function, we know it already containts some parameters like scale, delta's, qdamp, etc. I searched to see what exactly is there inside "addVar" in Github, but it is like a vicious circle, one code refers to the other and that one refers to the first one without saying what exactly and mathematically they are doing.

 I hope I could let you know what is going on in my mind and thanks in advance for your time and consideration.

 Any help or guidance would be highly appreciated. 
Reply all
Reply to author
0 new messages