Skip to main content

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
[stem-dev] Fw: Parameter estimation problem

forwarding for STEM-Dev record

----- Forwarded by James Kaufman/Almaden/IBM on 10/09/2017 11:52 AM -----

From:        jk
To:        Emily Nixon
Date:        10/09/2017 11:52 AM
Subject:        Re: FW: Parameter estimation problem

Hi Emily,

I downloaded the latest integration build of STEM, installed the latest Java 8 update, and tested the integration build on my mac.
The known bug re: the eclipse launder on mac is still unresolved to I opened the package contents, command-click on STEM, and I was able to run stem successfully.
After stopping and restarting STEM, I installed into the new workspace the Global Geographic Models, the Mexico USA pandemic flu scenario, and the Automated Experiment Example

We did migrate to a new build server so I wanted to check the build of STEMs build in geographic data is still ok. I could view the global models in the designer perspective. I could run the USA/Mexico scenarios,
and I could run the AutomatedExperiment scenario. I then opened the Experiments folder and I was able to run the AutomaticExperiment.automaticexperiment target.

In the error log you sent I'm seeing the message
        Unable to load content for resource platform:/plugin/
this is an unusual exception. Our built in data for countries uses the three letter country code so I would expect the letters STE to represent the three letter code for a country but there is no country with the code STE, and STEM has not built in data with that code.
The usual syntax by the convention (iso 3166)  used in STEM (e.g., for the united states) is:

USA_0 indicates the country USA at admin level 0 where 0 is the country level code
USA_1 indicates the country USA at admin level 1 where 1 is the STATE level code
USA_2 indicates the country USA at admin level 2 where 1 is the COUNTY level code

STEM allows users to build a graph with containment edges (eg, counties inside states, inside countries) in which case the computation is only computed at the lowest administrative level (eg counties in that example) and results can roll up the hierarchy.

Could you please zip up the project you created (in your workspace folder) and send it to us? I suspect you created a custom graph so I'm not sure why STEM is looking in the /data/country path for data. The _:_ symbol seems to indicate a problem.
Perhaps you created a flat square latttice graph  ?

fyi I also tested the command line options for running STEM headless and that also worked fine for me on my mac. Perhaps you could join a future STEM call and we could help to sort out that issue as well.

Back to the top