Open jakobj opened 2 years ago
Using a MSE error function can cause GP to prefer solutions which have wrong functional form but, for example, are close to the mean of the target data over solutions which have the correct form, but are shifted by a constant.
This problem can be easily circumvented by using scaled symbolic regression (https://link.springer.com/article/10.1023/B:GENP.0000030195.77571.f9). It's easy to implement, we should provide an example.
Using a MSE error function can cause GP to prefer solutions which have wrong functional form but, for example, are close to the mean of the target data over solutions which have the correct form, but are shifted by a constant.
This problem can be easily circumvented by using scaled symbolic regression (https://link.springer.com/article/10.1023/B:GENP.0000030195.77571.f9). It's easy to implement, we should provide an example.