Abstract:
This paper addresses to the problem of the point and the interval predictions of the time series on basis of recurrent first degree splines with depth 1. The conditions of optimality for coefficients of the calculated scheme are found. The results of numerical experiments for analytically given functions with errors are presented. A comparison with classical methods of prediction of the time series is given.
Key words:
spline, recurrence algorithm, optimization, prediction, time series.
Citation:
B. M. Shumilov, E. A. Esharov, N. K. Arkabaev, “Construction and optimization of predictions on the basis of first degree recurrent splines”, Sib. Zh. Vychisl. Mat., 13:2 (2010), 227–241; Num. Anal. Appl., 3:2 (2010), 186–198