Title: Interval prediction of oscillating time series based on grey system modelling

Authors: Gaofei Xu; Xiaohui Wang; Zhigang Li; Yang Zhao

Addresses: State Key Laboratory of Robotics, Shenyang Institute of Automation, Chinese Academy of Sciences, Shenyang 110016, China ' State Key Laboratory of Robotics, Shenyang Institute of Automation, Chinese Academy of Sciences, Shenyang 110016, China ' State Key Laboratory of Robotics, Shenyang Institute of Automation, Chinese Academy of Sciences, Shenyang 110016, China ' State Key Laboratory of Robotics, Shenyang Institute of Automation, Chinese Academy of Sciences, Shenyang 110016, China

Abstract: This paper presents an interval prediction algorithm based on grey system modelling, which is proposed for the forecasting of strong-oscillation time series with small samples. In the proposed algorithm, the upper and lower envelope of an oscillating sequence is obtained through cubic spline interpolation, and distance between the envelope and the fitted sequence derived from grey system model is dynamically expanded according to the oscillation intensity. After that, prediction value of the envelope distance sequence is calculated, and adjusted adaptively based on the new information priority principle. Finally, the interval prediction result is obtained. To verify the performance of the algorithm, five application cases from different fields were adopted. Compared with five representative algorithms in the recently related field, the proposed algorithm has distinct advantages in the prediction of small-sample strong oscillation time series.

Keywords: time series forecasting; interval prediction; oscillating time series; small samples learning; spline interpolation; grey system modelling; underwater vehicle; model residual prediction.

DOI: 10.1504/IJMIC.2019.104373

International Journal of Modelling, Identification and Control, 2019 Vol.33 No.2, pp.138 - 151

Received: 21 Mar 2019
Accepted: 01 Aug 2019
Published online: 06 Jan 2020 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article