Abstract
In the process of online prediction of multivariable non-stationary time series by kernel extreme learning machine (KELM), the dynamic characteristics of the system which are difficult to determine have always posed a big problem. We propose an online sequential prediction model with an adaptive forgetting factor (AFF) for multivariable time series to solve this problem. The multivariable time series instead of variable itself is reconstructed firstly. AFF is introduced into the objective function and can be adjusted iteratively and adaptively with the system changes. As a result, higher weight can be allocated for the fresh and more important samples while the old failure samples can be quickly forgotten. The model sparsification uses a fast leave-one-out cross-validation (FLOO-CV) method to set a prediction error threshold so that samples can be selected conditionally to form a dictionary. Besides, the dictionary parameters, including AFF and kernel parameters, are recursively updated simultaneously without increasing calculation complexity. The experimental results show that, compared with four fashionable KLEM methods, the proposed AFF-OSKELM has a better dynamic tracking ability and adaptability. Moreover, compared with single variable prediction, the spatial reconstructed multivariable has higher prediction accuracy and stability.
Highlights
Online prediction of non-stationary chaotic time series is an important research direction in the field of science and engineering
We propose a new multivariable time series prediction model with adaptive forgetting factor (AFF), namely AFF-OSKELM, to enhance the dynamic tracking capability and improve prediction performance and timeliness of non-stationary time series
The experimental results show that: (1) The proposed method has low computational complexity; (2) The AFF-OSKELM method has higher prediction accuracy and stability than other kernel extreme learning machine (KELM) based methods; (3) In the case of multivariable time series, the prediction effect of input multi-variable is better than that based on a single variable
Summary
Online prediction of non-stationary chaotic time series is an important research direction in the field of science and engineering. ELM achieves a balance between training error and model complexity by adjusting the number of hidden layer neurons to achieve optimal generalization performance. Most of the neural network models used in time series prediction are based on gradient descent learning algorithm, which has some defects such as slow convergence speed and easy to fall into local optimum. For this reason, ELM is proposed to determine the output weight by linear regression. The number of hidden layer nodes, which is an important parameter of ELM crucial to the performance of prediction model, usually should be selected by some time-consuming methods according to the learning tasks.
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
