3 ms·
Simple models have fewer sources of error, and the small number of parameters can be better validated against reality.
by HPsquared 6y ago
Simple models have fewer sources of error, and the small number of parameters can be better validated against reality.
- rramadass 6y agoThat certainly sounds plausible, however i feel that it is the applicability/non-applicability of parameters themselves that is the key. You need to choose the right parameters before even bothering about the model function and this is where prior knowledge and experience are important. The works of Gary Klein and Nassim Taleb are also relevant here.
- levosmetalo 6y agoNassim Taleb actually cites Prof. Gigerenzer in the topic of simple heuristics.