At present, the research methods of network pharmacology can be roughly divided into two categories: one is to establish a network prediction model of specific drug action mechanism according to the open database and published data, predict the drug action target, and analyze the drug action mechanism from the perspective of biological network balance. For example, CuJ[ 12] and others successfully predicted several components with anti-type 2 diabetes, such as rhein dione A, rhein dione C and sennoside C, by using virtual screening and network prediction technology.

The research of network pharmacology can not be separated from the construction of biological network. Biological network abstracts all kinds of substances and their interactions in organisms with the help of graph theory in the field of mathematics and complex network research methods.

Commonly used network topology analysis includes degree, betweenness, shortest path, bottleneck, central node, module, wedge point and so on. In biological networks, most nodes have small degrees, while a few nodes with large degrees are considered as central nodes. Median is the ratio of the number of paths passing through a node to the total number of shortest paths in the network.