Replication of machine learning experiment

Hello guys,

one big problem in science and machine learning seems to be the replication of experiments in machine learning. For example to analyse the recent google flu data problem:

http://www.sciencemag.org/content/343/6176/1203.full#xref-ref-13-1

Is there any scientific research done in this field how to actually replicate data science experiments in a meaningful way?

Would appreciate papers/keywords!