Deep Prediction Model Based on Dual Decomposition with Entropy and Frequency Statistics for Nonstationary Time Series.

Zhigang Shi,Xuebo Jin,Tingli Su,Xiaoyi Wang,Yuting Bai,Jianlei Kong

doi:10.3390/e24030360

Abstract

The prediction of time series is of great significance for rational planning and risk prevention. However, time series data in various natural and artificial systems are nonstationary and complex, which makes them difficult to predict. An improved deep prediction method is proposed herein based on the dual variational mode decomposition of a nonstationary time series. First, criteria were determined based on information entropy and frequency statistics to determine the quantity of components in the variational mode decomposition, including the number of subsequences and the conditions for dual decomposition. Second, a deep prediction model was built for the subsequences obtained after the dual decomposition. Third, a general framework was proposed to integrate the data decomposition and deep prediction models. The method was verified on practical time series data with some contrast methods. The results show that it performed better than single deep network and traditional decomposition methods. The proposed method can effectively extract the characteristics of a nonstationary time series and obtain reliable prediction results.

Highlights

We mainly investigated deep networks with data feature extraction for nonstationary time series, which is significantly different from previous studies
The decomposed intrinsic mode functions (IMFs) were predicted with gated recurrent unit (GRU) and the fusion result was obtained
The results show that the dual decomposition model can fully mine the data features and has a good predictive ability for nonstationary time series

Summary

Introduction

A time series is a significant representation of various objects and systems that describes their changing processes. The prediction of time series aims at estimating their future trends with hidden characteristics in the historical data. Time series prediction has gained widespread attention in many fields, such as meteorology [1,2], the stock market [3], environment pollution control [4,5], and data mining on the Internet. The reliable prediction of future trends can help administrators in comprehensive and scientific decision making [6]. At the core of making predictions is an appropriate model based on the features of the data. Predicting a time series effectively has been a hot issue in the fields of data mining and machine learning

Methods

Results

Discussion

Conclusion