I conducted an experiment which led me to believe that two sensors were time correlated somehow. Their signals do not show any obvious correlation, however their spectrograms show a strong similarity in a certain band of frequencies, with some harmonics in higher frequencies.
I am trying to use neural networks to learn the relationship. Namely, 1DCNNs. However there seems to be very little documentation on this exact approach online, so I've mostly been stabbing in the dark as to how many kernels, how deep to make the network, what size kernels, what layers to connect the 1DCNNs to, etc. Is there a common approach for problems of these types?