Wireless Social: Efficiency Information Spreading

Wireless social: A social network is a social structure made up of a set of social actors and a set of the dyadic ties between these actors (from Wikipedia), while in the wireless social network, the actors are movable. The mobile feature of wireless social network is one of the key factors for information diffusion. For example, in a university campus, students are usually in dormitories while they move to teaching buildings or lab buildings in the daytime. Consequently, the pattern of information diffusion at night and in the daytime is different.

At night:

    Mobile users in dormitories

    Mobile users in teaching buildings or lab buildings

In the daytime:

    Mobile users in dormitories
    Mobile users in teaching buildings or lab buildings

In order to make information spread efficiently, based on a community approach, we can choose some information seeds that are directly connected to the source of information and initiate the propagation of information in the network.

We identify the communities on the basis of two measurements: a) connectivity density: the connections among the nodes; b) distance density: the distance between a particular pair of nodes. Based on the detected community, we can find out the appropriate seeds. The nodes inside the community can act as the seeds, and we call them sub-source spreaders (green and red nodes). If communities overlap with each other, we can just select the nodes in the overlapped area as the seeds. This kind of nodes is named as ultimate-source spreaders (yellow nodes).

When a node inside the community originates a message, other inside nodes will receive the message with high probability. Consequently, there will be many nodes forwarding the message outside the community, and information will propagate efficiently. Then the entirety of the community can be viewed as a source. If communities overlap with each other, we just select the ultimate-source spreaders m. Then nodes in   and can easily receive messages from m. In this case, both the communities can serve as double sub-sources. With the data from Sina Weibo we have verified that if we select the sub-source spreaders or the ultimate-source spreaders as the seeds, the information will propagate efficiently. Further, the ultimate-source spreaders have a higher efficiency than the sub-source spreaders in information diffusion.