Hello, My question is probably going to seems evident for statisticians but not for me. If this newsgroup is not the right place to make this post, thanks to tell me.
I fabricate some systems that involve several parts assembled between each other. I want to determine the impact of the uncertainties on the parts (dimensions, positions ...) on the overall system response (through for example, the use of Monte-Carlo analysis) Here are the questions : #1 I measure some data (like thickness) of several parts. I would like to determine the best probability distribution function that fit my measurements. How to determine the needed number of samples ? How to choose the PDF's to be tested ? How to evaluate the accuracy of the result ? #2 I saw some software like Bestfit that seems to do this job. Is there any other one (free one would be prefered but commercial could be interesting as well). #3 Let's say that I have determined all the PDF of my parts. One of the characteritic C of my system is a function of several other variables X1, X2...XN, each having is own PDF. How to deduce the PDF of C ? Thanks a lot, Franck . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
