A fractal interpolation approach to improve neural network predictions for difficult time series data
作者：
Highlights：
• Fractal interpolation is used to increase fine-grainedness of time series data.
• A LSTM neural network is used to predict fractal interpolated time series data.
• All data are analyzed for chaoticity using the spectrum of Lyapunov exponents.
• All data are analyzed for persistency using the Hurst exponent and fractal dimension.
• Fractal interpolated approach outperforms regular approach.
摘要
•Fractal interpolation is used to increase fine-grainedness of time series data.•A LSTM neural network is used to predict fractal interpolated time series data.•All data are analyzed for chaoticity using the spectrum of Lyapunov exponents.•All data are analyzed for persistency using the Hurst exponent and fractal dimension.•Fractal interpolated approach outperforms regular approach.
论文关键词：Fractal dimension,Hurst exponent,Chaos,Hyperchaos,Lyapunov exponent,Time series data,Time series prediction,Deep learning,Machine learning,LSTM,Time series analysis,R/S analysis,Linear interpolation
论文评审过程：Received 17 July 2020, Revised 18 August 2020, Accepted 6 December 2020, Available online 13 December 2020, Version of Record 15 December 2020.
论文官网地址：https://doi.org/10.1016/j.eswa.2020.114474
