Both these models seem to be combinations of several models. The basic objective to go for these models is to improve the performance of the individual models. Now, how to differentiate them?
In hybrid modelling, different modelling technologies are combined. For example, a data-driven model may be combined with a theoretically derived one. This may be done for convenience. If good theoretical models exist for parts of the modelled process, but they are unavailable or too computationally intensive for other parts, the gaps between reasonable theoretical models can be bridged with data-driven ones.
In contrast, ensemble modelling combines several very similar models, with slightly perturbed parameters and/or boundary conditions. This is typically done to estimate the model sensitivity, improve the mean prediction through averaging, and/or estimate the range of possibilities.
In case of ensemble models, various methods are used and it can vote the desired outcome while in case of hybrid models different models are combined to get a better desired outcome.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Graphs