Japanese page

Connecting some points with line produces a design like a net. The design is called network. Point and line in networks are called node and link, respectively. There are some networks where we can easily know the structures because the structures are previously formed, and there are other networks where we can manage to know the structures by detailed observation and analysis. For example, there are many personal computers (PCs) placed here and there in a room. To make the exchange of data among PCs these PCs are connected by cables. In this case we can directly recognize the connectivity as a network, as PCs and cables are visible. On the other hand, when the exchange of data is made by air we cannot see the radio wave. We can only find PCs placed in various places and cannot recognize the connectivity among PCs. Hence, we cannot recognize the network at all in this situation - although we may assume that physical proximity plays a role. However, if there is the information about the exchange among PCs, we can know the connectivity among PCs and draw the network. The main research object is networks whose structures are unknown previously like the latter example. Networks are constructed by the identified connectivity based on the detailed data investigation. We apply the method of surrogate data and statistical models to construct the networks from data which reflect characteristics of the data and the meaning of the data relationship more accurately.