05-16-2018 05:34 PM
I'm using the log-normal cdf function to plot density curves overlaid on histogram plots of my data (mainly so I can examine how well dbns fit outliers). I'm running proc univariate to generate the scale and shape parameters for the fitted log-normal curve, then inputting the scale & shape values into the log-normal cdf function. However the resulting log-normal plots (using proc sgplot) are off - they don't match the default plots in proc univariate.
Do I need to transform the scale & shape parameters before I plug them into the cdf function?
05-16-2018 06:52 PM
Code? Since there are number of different plots available in Univariate we really need to see the code to see what your requested.
Example data? Either your data as a data step or use one of the SAS supplied data sets like SASHELP.CARS or SASHELP.BASEBALL to demonstrate the code and results.
And the SGPLOT code as well as how you build the set that SGPLOT uses.
05-17-2018 10:33 AM
Problem solved Ack! - I forgot to enter the "x" in the cdf('LOGNORNMAL', x, scale, shape) function!