简介：

Link prediction can help to uncover unknown relationships within a network by analyzing its structure. For certain networks, that information can then be used to speed up traffic flow, combat sexually transmitted disease, even find new disease drugs. In a regular social network, two people who share a friend are likely to themselves be friends with a direct connection. Adding this third connection creates a triangle – this is the structure on which existing link prediction algorithms focus. Conversely, our previous studies have focused on tree-like networks, which have many branches but very few cross links between branches. And very few triangles. Furthermore, we have found that in some other sufficiently sparse networks, such as sexual contact networks and water distribution networks, many long circle-like structures exist (that is large loops of interconnection). For example, if a person has a sexual contact with a partner, another secondary partner will typically have no sexual contact with the first person (few closed triangles). But sexual contact behaviors may occur in a one-by-one way with a number of persons, these then naturally form a long circle structure. Our novel algorithms to detect unknown links perform much better for long-circle-like networks than previous well-known algorithms, and may potentially help us to understand how to help stop transmitted disease, such as AIDS, COVID-19 and so on.

全文链接：https://doi.org/10.1103/PhysRevE.105.024311

免费的修改稿链接：https://www.researchgate.net/publication/358787216_Link_prediction_for_long-circle-like_networks