A comparison of two methods for modeling large-scale data from time series as complex networks

In this paper, we compare two methods of mapping time series data to complex networks based on correlation coefficient and distance, respectively. These methods make use of two different physical aspects of large-scale data. We find that the method based on correlation coefficient cannot distinguish the randomness of a chaotic series from a purely random series, and it cannot express the certainty of chaos. The method based on distance can express the certainty of a chaotic series and can distinguish a chaotic series from a random series easily. Therefore, the distance method can be helpful in analyzing chaotic systems and random systems. We have also discussed the effectiveness of the distance method with noisy data.