Data Investigation Distribution Analysis Tool

[tlewis]

I am getting an error when running the Data Investigation Distribution Analysis Tool. Will this work with a Double data type?

My error reads "error in fitdist(the.data,"norm",method="mle")

Perhaps my data is too small? It ranges from around 0.000044 to 0.000116.

[TrevorS]

Hello @tlewis

Can you please include a copy of your workflow and some sample data?

This will help the Community to troubleshoot your issue for you!

Thanks,
TrevorS

[PhilipMannering · Accepted solution]

Hi @tlewis ,

Having recreated the error I think it is because you're values are too small. You could try using a formula tool to multiply all your data by, say, 10000. This should do the trick. The goodness of fit statistics should be the same. The distribution parameters may need dividing by 10000 afterwards.

[CharlieS]

I did find that the Distribution Analysis tool has trouble with small values. In the attached example, I generate 100 random numbers [0,1]. The Distribution Analysis tool works fine on those values, but if I divide those values by 10000, the tool errors.

Perhaps there are limitations documented in the R packages used by this tool: "MASS", "fitdistrplus", and "stats".