SW-KRLS (sliding-window kernel recursive least squares)

The sliding-window kernel recursive least squares (SW-KRLS) is a model proposed by Van Vaerenbergh et al. [1].

  • SW-KRLS is the SW-KRLS model.

  • GridSearch_AllDatasets is the file to perform a grid search for all datasets and store the best hyper-parameters.

  • Runtime_AllDatasets perform 30 simulations for each dataset and compute the mean runtime and the standard deviation.

  • MackeyGlass is the script to prepare the Mackey-Glass time series, perform simulations, compute the results and plot the graphics.

  • Nonlinear is the script to prepare the nonlinear dynamic system identification time series, perform simulations, compute the results and plot the graphics.

  • LorenzAttractor is the script to prepare the Lorenz Attractor time series, perform simulations, compute the results and plot the graphics.

[1] S. Van Vaerenbergh, J. Via, I. Santamar ́ıa, A sliding-window kernel rls algorithm and its application to nonlinear channel identification, in: 2006 IEEE International Conference on Acoustics Speech and Signal Processing Proceedings, Vol. 5, IEEE, 2006, pp. V–V. doi:https://doi.org/10.801109/ICASSP.2006.1661394