How do you get a logic relationship between the input/output space dimension and the number of hidden neurons and/or the hidden layers in a problem to be solved by a MPL?
If you know any articles with something related with this or if you can imagine any logical relationship, please share with me and others the information.