Cancel
Showing results for 
Search instead for 
Did you mean: 

Re: Different model behaviour in version 11 and 10.1

Siemens Phenom Siemens Phenom
Siemens Phenom

Hi Aedel,

 

as you mentioned the sample of observations is typical for simulation studies.

Also the Box Plot is typical.

The 95% confidence interval is [0.6754 , 0.6968]. I would be satisfied.

 

Boxplot.gif

Such an outlier can happen. Maybe perform a simulation study with more observations.

I believe that there are even smaller values.

 

To your last question: The differences between the generation of the observation is

  • Version 10.1: Change the seed table: menu Tools > Random number seed values
  • Version 11.0: Change the Random Number Seed Variant (Eventcontroller menu Tools)

 

Regards,

Peter

Re: Different model behaviour in version 11 and 10.1

Experimenter
Experimenter

Hey Peter,

 

Thanks a lot!

If the confidence interval was around 65% I'd be really happy Smiley Happy

That's the value that comes from the data of the real system.

 

While I was building the model, I only used single runs to see, how it's performing. Now it looks like I built it around an outliner.

 

I'll try an experiment with more runs, but I guess, it'll come out around 68% again.

 

Have a nice evening!