| dc.description.abstract | The amount of available information on the web is increasing, and companies are expanding the way to both collect and use the information available. This is the situation for both personal information, and technological information such as HTML-documents. Throughout this paper, I will describe the development of a semantic web recommender system that aims to reduce the amount of personal information needed to provide personal web recommendations. Semantically marked up documents on the web contain information, which is not necessarily provided in a user interface. This means there are possibilities to expand the area of use for this technology. The use of Semantic Web-technologies can therefore contribute to reduce the need of giving away personal information on the web. This thesis is divided in two parts: The first part focuses on the development of a semantic application, and the new area of use of this technology. The other part focuses on how standard recommenders handle privacy concerns on the web. The thesis will provide a description of the development of the recommender system, as well as an explanation of online privacy and how different web service providers' deals with it. The system uses an RDFa-API to collect semantic information available on web-documents, and further uses this information to provide recommendations for the users. This thesis concludes that it is possible to recommend new web content for a user with this method, but the collected information varies wildly. This is related to both the complexity of the developed system and the way things" are marked on the web. It is further shown that this method can reduce personal information, however it is shown that users who are comfortable with social medias are not worried about privacy on the web. | en_US |
