The training objective is used to find the optimal neural network structure as well as selecting the best inputs and/or parameters for the inputs. Click on the arrow to view the entire list. One of our favorites is on top, but you’re encouraged to experiment. Some of the objectives create similar models, but that isn’t an exact science.
There is a default value for the number of hidden neurons in the model. The default value is set on the low side to encourage more robust models. As you add hidden neurons you increase the possibility of overfitting the model to the training data. Note that it’s more important to find good inputs than to adjust hidden neurons.
If you’re trying to build a prediction for a trending market, you might want to experiment with checking this box.