The development of the Internet and the World Wide Web has changed our outlook on the world. No longer is information only given in book form, with items catalogued in an orderly manner. More and more often, information is presented in a Web-like form, as a mass of items with interconnecting links. We refer to such a collection as a networked information space.
In most cases the links, and to a lesser extent the information, evolve rapidly over time. For example, in the World Wide Web, pages and links are constantly being added, updated or deleted. In networks representing the exchange of information, such as those representing phone calls made or emails sent, or traffic routed through the Internet, the dynamic behaviour of the links is an essential feature.
The practical importance of modeling and mining dynamic networked information spaces is wideranging. Phone-call and email networks are of interest to law-enforcement agencies, because mining such graphs can lead to the detection of unlawful activities. Classifying Internet traffic flow data according to the applications that generated it is of great importance to network management. The analysis of the dynamic evolution of networks formed by online discussion groups can give insight in socio-economic factors such as the spread of influence and the concentration of authority. On the Web, good models of dynamic growth have the potential to improve the ranking of search engine results.
A crucial component of the link analysis of networked information spaces is the identification of the community structure. The typical link structure of a networked information space has the form of "lumps" of locally dense structures, floating in a sparse "soup" of loosely connected nodes. The lumps correspond to communities of items that are somehow related: Web pages on a related topic, emails between members of the same research group, phone calls between subscribers in the same town. The challenge of the MoMiNIS project is to extract community structure and follow its evolution over time, and use our findings to understand and improve the wired world we live in.
