A Model for Automatic Generation of Multi-partite Graphs from Arbitrary Data

In this paper we propose a generic model to generate basic multi-partite graphs obtained by associations found in arbitrary data. The interest of such a model is to be the formal basis behind a tool for automatic graph generation that we are developing. This tool will automatically generate the basic multi-partite graphs that represents the arbitrary data provided as input. We consider input data as collections of complex objects composed by a set or a list of heterogeneous elements. Our tool will provide an interface for the user to specify the kind of nodes that are relevant for the application domain in each case. Those nodes will be obtained from the complex input objects by simple extraction rules. The objective of this paper is to present the model to represent basic multi-partite graphs and the way to define the nodes of the graph using simple derivation rules defined by the user. In order to validate our model we give three examples of radically different data sets. Those examples come from the Web log queries, processing text collections, and bibliographic databases.