Interaktive Angleichung als Modell für die Mensch-Computer-Interaktion im Semantic Web

Currently, the WWW includes more than 23 billion Web pages and this trend is rising. The amount of information contained in these websites is enormous and a suitable access to it therefore of great interest. Most information in the WWW, however, exists only in the form of natural language texts, images or videos; embedded in HTML documents. Due to the ambiguity of these contents, a purely automatic search is often limited. Thus, for example, texts can often only be searched based on their syntax but not their semantics. This, however, represents an important limitation for the access to information in the WWW. To address this problem, the Semantic Web was proposed as an extension of the WWW. All information in this extension is provided with a clear meaning, which can be interpreted by computers. This allows to automatically search for information based on its semantic and thus to improve the quality of information access in the Web considerably. However, it requires the semantic of all the information in the Web to be available in explicit form, namely as semantic data. The problem is that, on the one hand, automatic methods for extracting semantic data from existing websites have a satisfactory quality only in very few cases; on the other, purely manually created semantic data exist only for a small fraction of the available information in the Web, because the creation is very time-consuming. For this reason, only a small percentage of the information that is available in the WWW is already represented in the Semantic Web. Another problem, besides the di culty to create semantic data, is the di culty to access information in the Semantic Web. By using conventional approaches, users are often not able to clearly de ne the meaning of what they are looking