Difference between revisions of "Add Problem Content"
(→Notes/Comments) |
|||
Line 40: | Line 40: | ||
<br> | <br> | ||
---- | ---- | ||
− | See | + | See completed [[DataShop 3.x Features]]<br> |
+ | See on-going [[DataShop 4.x Features]]<br> | ||
+ | See prioritized [[DataShop Feature Wish List]]<br> | ||
+ | See unordered [[Collected User Requests]] | ||
[[Category:Protected]] | [[Category:Protected]] | ||
[[Category:DataShop]] | [[Category:DataShop]] | ||
+ | [[Category:DataShop Feature]] |
Latest revision as of 18:48, 29 October 2010
Status: Prioritization Needed
User Story
As a researcher, I want to get problem content so that I can better predict performance (e.g., in the KDD Cup task) especially when there is not existing strong KC model.
Notes/Comments
- Email from Ken to Ryan, May 2010
I'd like to get your method for extracting the problem descriptions from CL files in use as part of DataShop -- that is, to produce files that can be attached to associated datasets that provide the problem text information.
- Response from Ryan, May 2010
It's not that easy, unfortunately. :-| Thinking back, I think I must have had Jay step through problems rather than using a problem dump (and maybe he just looked at the first one? not clear). When I said I dumped this out via code, I think what I was thinking of was the hint messages, which I dumped out of DataShop data. But it *is* doable, just a good bit of work. :-) Go to C:\Program Files\Carnegie Learning\2006\Administrative\Curricula\Unit Data (or similar folder) Find the lesson of interest Search for "problem id". Find the problems. C:\Program Files\Carnegie Learning\2006\Administrative\Curricula\Problem Data Rename problem-data.jar to problem-data.zip, then extract it (do this on a copy) Then go into Initial State Rename initial-state.jar to initial-state.zip, then extract it Find the problem id.tre (e.g. if you are looking for problem Sys02, find sys02.tre) Search in that file for <<Scenario>> The scenario text follows (after some meta-data). No doubt, the scenario text could be extracted from this, either via code, or undergraduate labor. Not trivial, though.
See completed DataShop 3.x Features
See on-going DataShop 4.x Features
See prioritized DataShop Feature Wish List
See unordered Collected User Requests