I have a neural network that takes 1935 inputs, so I'm wondering if there is a general rule for how many layers the network should be.
Should the number of neurons be descending by a certain amount?
$\begingroup$
$\endgroup$
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
It doesn't depend on the number of inputs directly. Your network has to be large enough to capture the structure of the problem and the data. This process would involve some trial and error experimentation
For challenging predictive modeling problems, deep neural networks may be a heuristic approach.
For the number of neurons and how much they should descend too, use trial and error. We cannot specify the “best” number of neurons analytically. We must test.
Exploring and implementing more such models would help you develop better intuition about these, still involving some trial and error.
