Issue with Sigmoidnet function input

Can someone please post what sigmoidnet function, under nonlinear ARX model in the system identification toolbox, is taking as input and how it is taking a time series data?

More precisely, the function of the sigmoidnet is written below,

F(x) = (x-r)*P*L + a_1 f((x-r)*Q*b_1+c_1) + .. ... + a_n f((x-r)*Q*b_n+c_n) + d

where f is the sigmoid function and n is the number of units of the function.

But, in the above formulation, what is x? and how the regressor mean (r) is calculated?

Answer 1

[Roshni Garnayak]

x is an m-dimensional row vector representing the regressors and r is the mean of the regressors x. To understand the structure of nonlinear ARX models, refer to pages 11-18 to 11-19 of the following documentation:

https://www.mathworks.com/help/pdf_doc/ident/ident.pdf

Refer to pages 11-63 to 11-64 of the above documentation for an example that preforms a low-level computation of the nonlinearity response for the “sigmoidnet” network function.