Continuous-time system identification for discrete data by curve fitting

We propose continuous-time system identification using discrete data and continuous curves of the data obtained by curve fitting. We describe a method for fitting a finite Fourier series to discrete data and show that the fitted curve can be created as the output from a bilinear neural network system. By using this method for fitting, we propose a method for determining the optimal values for coefficients in the differential equation which expresses a continuous-time system.