I have a dataset of observed data and I want to test if they are ~ LogNormal (μ;σ)
I have used the following:
Proc Univariate data=myData;
var Var1;
histogram / LogNormal;
run;
I have, as an output, different tables and I would like to store in a Table the results of the "Parameters for Lognormal Distribution".
Parameter | Symbol | Estimate |
---|---|---|
Threshold | Theta | 0 |
Scale | Zeta | 9.09 |
Shape | Sigma | 1.04 |
However, it seems to me that I can only store basics statistics (e.g. normality test, skewness test)
Anyone may help?
Thanks in advance!
Enable ODS Trace and you can find the table that stores this information, then output it with ods output
. For example:
ods trace on;
proc univariate data=sashelp.cars;
var horsepower;
histogram / lognormal;
run;
ods trace off;
Check the log and you'll see:
Output Added:
-------------
Name: ParameterEstimates
Label: Parameter Estimates
Template: base.univariate.FitParms
Path: Univariate.Horsepower.Histogram.Lognormal.ParameterEstimates
-------------
Use ods output
to grab the parameter estimates and save them to a table:
proc univariate data=sashelp.cars;
var horsepower;
histogram / lognormal;
ods output ParameterEstimates;
run;
VarName Histogram Distribution Parameter Symbol Estimate
Horsepower 1 Lognormal Threshold Theta 0
Horsepower 1 Lognormal Scale Zeta 5.321345
Horsepower 1 Lognormal Shape Sigma 0.328634
Horsepower 1 Lognormal Mean 216.0145
Horsepower 1 Lognormal Std Dev 72.95021