Statistics Problem - Generate Trial Data

**Greg777** · 05-10-2012, 06:53 PM

Hello,

I have a data set with a time series containing mean values, and values that are 1 standard deviation likely to occur on either side of the mean value (i.e. 34.1% likely). Figuring an answer for just one year's data is sufficient, since I will be able to extend the solution to the other years. I'd like to use a log-normal distribution to randomly generate trial values for each year (i.e. simulation, not an expected value). Splitting the log-normal distribution into a piece-wise function around the mean is OK. I have tried a number of approaches, but none have been promising; so I won't complicate the presentation with them. I've attached a workbook.

Any help appreciated!

Greg

**shg** · 05-10-2012, 07:14 PM

Here are two UDFs (RandLogNorm and RandLogNormMV) that generate log normal deviates, depending on your parameters.

Please Login or Register  to view this content.

**Greg777** · 05-10-2012, 07:40 PM

Thanks for the code. Unfortunately I'm not great with VB. I tested the custom function as RandLogNormMV(0,1) and it returned a #NAME? error. Would you please upload a copy of the example with the code functioning within it? I should be able to do the application of the formula from that point.