We present a model which creates a linear system of equations of symmetric and heptadiagonal. Therefore, for huge volume of data, the governing matrix is sparse and has a lot of zero entries, hence there is no need to save the entire values of matrix A. It is necessary to keep the bandwidth of data with the thickness of 4n. The non-zero entries part for a linear system with dimensions 100100 is about 4%. For testing the presented model, 4 problems with highly nonlinearity of data variations are selected. The results show the applicability, the efficiency and the simplicity of the model for approximation of highly non-linear data.
Keywords
Approximation, B-spline, Heptadiagonal Matrix, Bandwidth, Non-linear
User
Information