Nonuniversality of the horizontal visibility graph in inferring series periodicity

The filter horizontal visibility graph (fHVg) algorithm was recently proposed to detect the hidden periodicity of intrinsically periodic series under the pollution of noise. In this work, we evaluate the reliability of this algorithm by taking into account the effect of finite size and noise pollution, and something intriguing is found. The fHVg is first applied to logistic map with period 2 and 3, and numerical results suggest that the accuracy of fHVg is not affected by the length of tested series. It is effective in analyzing very short time series but sensitive to extrinsic noises. However, the fHVg has unexpected limitations that lead to spurious results. It lacks generality and shows inability when applied to logistic map with period 4 and to the monthly mean temperature dataset from real-world. (C) 2019 Elsevier B.V. All rights reserved.