I have a problem where I have multiple wind speed forecasts and I am trying to assign each forecast a different weight to obtain the best “blend” forecast that gives me the best forecast performance. Assuming each wind speed forecast performs differently under external conditions such as temperature and humidity.

I understand this could be modeled as an optimization problem, but I am curious if this can be solved through machine learning since NN are basically solving optimization problems and are made up with weights and biases.

So given the forecasts, and possibly, given the external conditions (for example, temperature), the model returns the optimal weight selections.

Looking forward to hear what y’all think. Thanks in advance!