Abstract: Patent is an important embodiment of innovation.Many people will inquire about the process of patent application on the Internet and learn the application steps before patent application.In fact, the online searching is also a way to know whether innovative enterprises or individuals attach importance to innovation.This paper analyzes the dynamic characteristics of Baidu search index time series with the keywords of “patent application” from the perspective of a new time series analysis, namely from the perspective of network.The time series of Baidu search index is transformed into a complex network by using the principle of visibility graph algorithm, and its parameters are calculated to analyze the topological structure of the network.Firstly, by calculating the complex network, it can be found that the patent attention of each province has certain differences.Secondly, the study shows that most of the networks are scale-free networks and the original time series have fractal characteristics.Finally, by clustering, 31 provinces can be divided into 3 categories according to the characteristics of complex networks.This paper analyzes the data of Baidu search index from 2011 to 2018.By dividing the community structure, the time series period and the central node’s influence on the search index can be found.

Key words: Baidu index, Complex network, Concern, Patent application, Visibility graph algorithm