Composite Documents and Uncertain Inference

The composite nature of multimedia documents requires a more powerful knowledge repre sentation for indexing than the pure set of terms. Object-oriented data modeling i s a widely known and accepted approach to represent knowledge. [Meghini et al. 93] introduces the usage of object-oriented principles in combination with logic for improving informat ion retrieval. This workshop contribution presents ideas to combine uncertain inference with obje ctri nted modeling in order to achieve a suitable model for multimedia information retri eval. This work introduces the probabilistic extension of the model presented in [Rölleke & Fuhr 96]. Figure 1 depicts the major issues of the model. The documents d1 andd2 consists of words and sections. The square brackets indicate the composite (aggregated) structure of documents. This concept of aggregationallows for reflecting the composite nature of multimedia documents appropriately. In addition, it is suitable for modeling retrieval among distri bu ed environments as indicated by the two databases db1 anddb2. We consider databases, documents, and sections as contextswhich define a local frame for a logical program. A logical program is a set of facts and rules for defining a set of pr opositions. Following the object-oriented principles, we consider three types of proposit ions, namely classification, generalization, androles. Classification serves to group objects within certain classes (concepts). For example, document d1 states that object peter is an instance of class sailor. Generalization serves to define class hierarchies. For example, ever y picture is a document. Roles represent the relationships between objects. For example, ary is the author of d1. In addition, our model allows to use a predicate with arity zero (e. g. sailing). This corresponds to the classical set of terms approach and provides the fami liar way for describing the content of documents.