0
$\begingroup$

I have 60+ models (could be basic if-else rules or complex neural networks) that describe population risk. Each of the models has at least one output (often more). What is the technical term for the top-level model that summarises every risk by taking the 60+ model outputs as inputs? I don't want a name for the whole process ;)

I'm thinking 'summary' or 'integrated' model, but if there's an already existing term, that would be great.

$\endgroup$
  • 2
    $\begingroup$ ensemble would be my guess $\endgroup$ – probabilityislogic May 24 '19 at 3:49
  • $\begingroup$ @probabilityislogic Yess perfect - thanks for that. I'll accept that if you make it an answer :) $\endgroup$ – Josh May 24 '19 at 4:30
0
$\begingroup$

The general term for such techniques is ensemble learning as suggested by @probabilityislogic. Specifically, if the level-1 model is also trained, it would be an example of Stacking.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.