Missing input: ExampleSet regardless IOMultiplier

Legacy User

HI,

I'm trying to compare two different learners with the T-Test operator. My model is pretty much
the same as sample 13_SignificanceTest. Even though I use the IOMultiplier operator to provide
the example set for both XValidation chains, RapidMiner issues an error telling me that an
example set is missing for the second learner. To reproduce, just run the mentioned example and
you get:

[Error] LinearRegression: LinearRegression: LinearRegression: Missing input: ExampleSet
[Error] There was 1 error.


So, is the error message redundant (since my example/the sample can be completely executed
and reasonable results are produced) or is there really something wrong?

Thx,
Benjamin

TobiasMalbrecht

Hi Benjamin,

the error appears, since the first [tt]XValidation[/tt] operator does not return an example set. Unfortunately, RM does not recognize that there is still one example set produced by the previous [tt]IOMultiplier[/tt]. That is already a known drawback of the RM process validation. The process validation does not consider the count of objects passed between operators, hence the error. Generally, you can ignore the error. An alternative which should not result in the error message would be to remove the [tt]IOMultiplier[/tt] and to set the parameter [tt]keep_example_set[/tt] of the first [tt]XValidation[/tt] operator to true.

Hope that helps,
Tobias