Marine multiple time series relevance discovery based on complex network

Abstract

Ocean measuring point is an important way to obtain many kinds of marine data. Reasonable layout of ocean measuring points can efficiently obtain marine data. At present, a marine measuring point can acquire multiple types of marine data, only by comprehensively using multiple types of ocean data we can more effectively discover the relationship between various ocean measuring points. This paper proposes a mapping method for fusion marine multiple time series into an image, and uses the similarity between different images to construct a complex network. Also, We build a complex network of marine multiple time series by selecting appropriate thresholds. Compared with the traditional method, the network constructed by our approach can find more accurate rules.