Ventana software support forum

hi, this question may be too simple for most people here.

for the target variable, the sensitivity grant could display mean value, 50%, 75%, and any other user defined confidence bounds. but I would like to save these information as a table, for example, I have interest to save the two lines (together they make up the 95% confidence bound) over time as a table. Could you help me on this?

thanks,

Yong

You cannot save the values in the tables. All you can do is export the sensitivity run and re-calculate them in something like Excel.

Not so. In DSS you can use the % keyword with SENS2FILE in a command script. This will get integrated into the normal export process at some point. https://www.vensim.com/documentation/in ... s2file.htm

thanks a lot to both of you.

So without DSS, all I can get is a visualization? we could roughly estimate the values of any confidence bounds but no way to the specific values? that is really a pity: visualization is important, but it seems to me, most people would like to obtain the information quantitatively, will be much easier to share and report to others. Why the VENSIM does not allow this? thanks!

Remember that PLE is a learning tool, not a production tool.

That said, we're upgrading most of the tools, and this is the kind of thing that will get implemented in the new system.

yongyang wrote:thanks a lot to both of you.

So without DSS, all I can get is a visualization? we could roughly estimate the values of any confidence bounds but no way to the specific values? that is really a pity: visualization is important, but it seems to me, most people would like to obtain the information quantitatively, will be much easier to share and report to others. Why the VENSIM does not allow this? thanks!

I think it's because analysing sensitivity runs is not as straightforward as looking at a single graph (from a single run). It's always been assumed that to do any proper analysis, you'd need the results in a spreadsheet.

I've added a new ticket to our system to implement this, so it should be done soon.

thanks a lot. good to know. What i am using is PRO not PLE.

tomfid wrote:Remember that PLE is a learning tool, not a production tool.

That said, we're upgrading most of the tools, and this is the kind of thing that will get implemented in the new system.

great! thanks a lot!

Administrator wrote:

yongyang wrote:thanks a lot to both of you.

So without DSS, all I can get is a visualization? we could roughly estimate the values of any confidence bounds but no way to the specific values? that is really a pity: visualization is important, but it seems to me, most people would like to obtain the information quantitatively, will be much easier to share and report to others. Why the VENSIM does not allow this? thanks!

I think it's because analysing sensitivity runs is not as straightforward as looking at a single graph (from a single run). It's always been assumed that to do any proper analysis, you'd need the results in a spreadsheet.

I've added a new ticket to our system to implement this, so it should be done soon.

Administrator wrote:

yongyang wrote:thanks a lot to both of you.

So without DSS, all I can get is a visualization? we could roughly estimate the values of any confidence bounds but no way to the specific values? that is really a pity: visualization is important, but it seems to me, most people would like to obtain the information quantitatively, will be much easier to share and report to others. Why the VENSIM does not allow this? thanks!

I think it's because analysing sensitivity runs is not as straightforward as looking at a single graph (from a single run). It's always been assumed that to do any proper analysis, you'd need the results in a spreadsheet.

I've added a new ticket to our system to implement this, so it should be done soon.

Hi, could you give me any ideas how soon this function could be added? of course, just a rough estimation, I asked this because if the timeline is within several weeks, I will wait. if within several months, I will try to figure out other ways for solve this problem. thanks a lot!

yongyang wrote:Hi, could you give me any ideas how soon this function could be added? of course, just a rough estimation, I asked this because if the timeline is within several weeks, I will wait. if within several months, I will try to figure out other ways for solve this problem. thanks a lot!

I cannot give you an estimate. The quickest thing for now would be to export the results file to Excel and do the analysis there.

Administrator wrote:

yongyang wrote:Hi, could you give me any ideas how soon this function could be added? of course, just a rough estimation, I asked this because if the timeline is within several weeks, I will wait. if within several months, I will try to figure out other ways for solve this problem. thanks a lot!

I cannot give you an estimate. The quickest thing for now would be to export the results file to Excel and do the analysis there.

Could you tell me how to export the results file to Excel? if it could be done, then problem solved. thanks a lot!

Open the Vensim help and search for exporting datasets. It's the first link in the results.

Administrator wrote:Open the Vensim help and search for exporting datasets. It's the first link in the results.

i see, so you mean i still have to export all simulation data and then compute the 95% confidence interval manually?

yongyang wrote:i see, so you mean i still have to export all simulation data and then compute the 95% confidence interval manually?

Yes. When you export a sensitivity run, all the simulations are exported.

Hi I would like to check that the confidence bounds used within the Sensitivity graph refers to the number of graphs that fall within the bound right? It does not refer to the confidence intervals used in statistical analysis right? Or am I missing some statistical knowledge that I should know here? I did a 1000 simulation runs and the range of possible data was huge, which is fine, but when I calculated confidence intervals at each time point it was tiny. Sorry if I sound silly here for doing something wrong here.

Oh nvm, just realized Vensim represents the graphs based on percentiles, which makes sense! Thanks!

Curious question, is it possible for me to buy a CD of the software and install it in a non-internet connected computer?