Help me create a program using Artificial Neural Network

Good evening,

help me, please, solve the problem. With neural networks I'm just starting to work, but I think, only it can help me solve the problem with many parameters.

There is a set of experimental points (x, y1-yn), each of them can be described by a curve of the form y = f (x, n1-nn, m, a, b, c ..), where n1-nn is the known vector for each curve y1-yn, m is an unknown common vector for all curves, a, b, c .. are unknown common parameters for all curves. I must find a common vector and common parameters for all the curves, so that the experimental points lie as close as possible to the corresponding functions. For example set {x, y1} can be fitted by y1=f (x, n1, m, a, b, c ..), {x, y2} by y1=f (x, n2, m, a, b, c ..)..

To solve this the problem, I select the Neural Network Fitting and for configuration of the neural network, I do this: I generated a data array in which I went through combinations of parameters n1-nn, m, a, b, c .. over a wide range. In the input, I passed a matrix with the values (x, y1-yn) in the form of strings. The matrix (m, a, b, c ..) is also output by strings. After the configuration, when I use the function sim (net, [x; y1-yn]), it produces a certain combination of numbers (m, a, b, c ..). My question is.

How to submit a data array to the sim function so that the network determines the general parameters(m, a, b, c ..) from the range of the values x1-xn and y1-yn? And do not find them at a specific point, because this solution does not make sense. And is it worth to specify x1-xn in the input parameters, if the function is periodic and experimental points can be shifted in phase?

If there are other ways to solve this problem, I will be grateful!

Thanks for answers!