A Multidimensional Approach to Characterizing and Visualizing Latent Relationships in Enterprise Social Networks
01 June 2012
Social networking applications, one hallmark of Web 2.0 [13], aim to make it easier to create, maintain, and understand social relationships [3]. Most social networking applications include relationship discovery features to help users connect with or "find" other users of interest to them, either by providing explicit recommendations or via tools that enable users to visualize their social network to discover other users of interest on their own. Consumer-level social networks typically recommend other users that a given user is likely to know already offline, perhaps through shared friends, shared attendance at a particular school, etc. One example of this is the "Friends You May Know" feature in Facebook*. Most commercial social networking applications, such as Facebook in the consumer space and Yammer* in the enterprise space, do not publicize the algorithms they use to make friend recommendations; indeed these algorithms are closely-held trade secrets. However, it is not unusual for users to complain that the recommendations made by these tools are often way off the mark.
