Simplification Algorithms for Topic Maps

Abstract

This thesis is about methods to tackle large quantities of information. It will give an introduction to how information is structured to increase understanding and how to present this in a simple and effective manner. The emphasis is on how to reduce all of the information which is browsable in an application to only that which is relevant for a reader.

We will give an introduction to a technology known as topic maps, which is a tool for organizing and navigating large information pools. They are very powerful when it comes to organizing information, but they tend to become very large. This presents a problem because it makes the navigation of the topic map more difficult.

The information inherent in a topic map, which can be characterized as a semantic network, can be visualized as a graph. This graph can be used to simplify the topic map itself. This will be done on the basis of the ab modeling language, a simple language that represents information models as directed graphs and which includes a set of operations to reduce a graph in order to reach simpler information representations. Because of the ab modeling language s behaviour perspective, the emphasis is on causal models of information, i.e how a change in one concept induces a change in another concept. Most of the methods themselves are based on these behaviour oriented models.