Approximation of data using non-integer harmonics series

作者：Tenreiro Machado J A; Lopes Antonio M

来源：Nonlinear Dynamics, 2017, 89(4): 2845-2854.

DOI：10.1007/s11071-017-3629-4

• Time series
• Fourier series
• Harmonics
• Curve fitting
• Forecasting

This paper introduces a method for approximating real world data by means of non-integer harmonics series (NIHS). The NIHS terms are sinusoidal functions of arbitrary amplitude, frequency and phase shift that can be computed by direct numerical algorithms. Applications go from obtaining time derivatives up to forecasting future values. Three illustrative examples with the Dow Jones Industrial stock market index, the Europe Brent Spot Price FOB and the daily temperature records of New York city are studied. The results show the effectiveness of the NIHS when dealing with real world time data series.

• 出版日期2017-9