Fractional Linear Prediction

Functions providing fractional linear prediction of one-dimensional signal.

This submission contains functions implementing the fractional linear prediction (FLP) models to estimate the one-dimensional signal. Two versions of FLP are implemented.

The first approach to FLP is using the whole history of the signal ("full" memory) - function flp_f.m. The second approach to FLP uses the "restricted" memory (restricted to two, three or four previous samples) - function flp_r.m.

The methods, models and applications implemented in the form of the proposed functions were presented in the works [1], [2]. The MATLAB implementation of the discretization of the fractional order derivative can be found in [3].

[1] Tomas Skovranek, Vladimir Despotovic: Signal prediction using fractional derivative models. In: Handbook of Fractional Calculus with Applications, Volume 8: Applications in Engineering, Life and Social Sciences, Part B, Pages 179-205. De Gruyter, 2019.
https://doi.org/10.1515/9783110571929-007

[2] Vladimir Despotovic, Tomas Skovranek, Zoran Peric: One-parameter fractional linear prediction, Computers & Electrical Engineering, vol. 69, July 2018, Pages 158-170 (Included in Special Issue on Signal Processing, March 2018).
https://doi.org/10.1016/j.compeleceng.2018.05.020

[3] Igor Podlubny, Tomas Skovranek, Blas M. Vinagre Jara: Matrix approach to discretization of ODEs and PDEs of arbitrary real order, MathWorks, Inc., Matlab Central File Exchage, 2008 (Updated 04 Mar 2016).
https://www.mathworks.com/matlabcentral/fileexchange/22071