Multi-scale Gaussian process experts for dynamic evolution prediction of complex systems
作者：
Highlights：
• A new model to predict dynamic evolution of complex systems is proposed.
• Intrinsic time-frequency-energy patterns of the systems is realized.
• Those intrinsic patterns capture the nonlinearity and nonstationary dynamics.
• This multi-scale Gaussian process model outperforms classical forecasting models.
摘要
•A new model to predict dynamic evolution of complex systems is proposed.•Intrinsic time-frequency-energy patterns of the systems is realized.•Those intrinsic patterns capture the nonlinearity and nonstationary dynamics.•This multi-scale Gaussian process model outperforms classical forecasting models.
论文关键词：Multi-scale Gaussian process,Intrinsic time-scale decomposition,Nonlinear,Nonstationary,Multi-step forecasting
论文评审过程：Received 20 July 2017, Revised 18 December 2017, Accepted 13 January 2018, Available online 31 January 2018, Version of Record 3 February 2018.
论文官网地址：https://doi.org/10.1016/j.eswa.2018.01.021
