Generate expected outputs as part of the generation process

ftsrg / trainbenchmark

The Train Benchmark framework for evaluating incremental model validation performance

https://ftsrg.mit.bme.hu/trainbenchmark/

Eclipse Public License 1.0

12 stars 19 forks source link

Generate expected outputs as part of the generation process #52

Open szarnyasg opened 8 years ago

szarnyasg commented 8 years ago

Related to #26.

szarnyasg commented 7 years ago

I've been thinking about this - the generator on its own cannot determine the expected output. Instead, it has to be executed on a tool. So, cross validating results or the results sizes (as it is done by the LDBC SNB benchmark) is a good way to do this.