Creating PDFs to Converting Data into Histograms and Fitting Curves

[member 428835]

hey all (again)!

how do people create pdf's? it seems like we start by collecting info and turning it into a histogram. then do we simply look at an array of curves, or is there something more, such as what is done of finding the least squares method for fitting data to a line?

thanks!

 

[FactChecker]

Most of the time the data is assumed to fit some known distribution with a theoretical PDF and parameters. Then the parameters are estimated. If you are creating a PDF from scratch from the data, I would make a CDF first. The CDF is an integral of the PDF so it is smoother and less dependent on the choice of histogram partitions. Then a curve can be fit to the CDF and the PDF can be estimated.

 

[member 428835]

Thanks!