Institute for Web Science and Technologies · Universität Koblenz - Landau
Institute WeST

Topics

Computational Social Science

CSS develops algorithms and novel, non-obtrusive methods for the social sciences based on state-of-the-art approaches from the domains of machine learning, data mining and network analysis. Current projects focus on the quantitative analysis of political electoral processes and dynamics based on social media data as well as the quantitative analysis of other social processes in social media or log data.

People

JProf. Dr. Claudia Wagner

Dr. Florian Lemmerich

Interactive Web and Human Computing

The objective of the working group Interactive Web and Human Computing is to make interaction with computers easier for Web users, in particular in situations where humans must deal with a huge amount of multimedia and knowledge content, such as given in the Web. Furthermore, the group investigates how human expertise can be used in the Web to enable intelligent problem solving that computers could not achieve on their own, because they lack relevant background knowledge or cognitive abilities that humans have. The outcome is social machines with previously unattainable collective intelligence. An example for a social machine is the definition of semantic relationships between concepts by crowd workers in the Web exploiting their collective intelligence.

People

Dr. Chandan Kumar

Prof. Dr. Steffen Staab

Semantic Web

The Semantic Web is a vision of a world-wide network of data (“linked data”) that allows for intelligent systems that search, retrieve, and aggregate useful information. In comparison to today’s Web, it goes beyond merely presenting the information content on web pages by also representing the meaning (semantics) of the information.

The focus of the working group Semantic Web lies in research and development of approaches to manage and discover semantic information. For this purpose, logic-based approaches are utilized for intelligent and robust reasoning with semantic data and the integration of data from multiple sources. This includes aspects of data management such as indexing, efficient query processing, and benchmarking. In today’s fast growing cloud oflinked data, these methods are needed to discover and analyse interesting and relevant information.

For this purpose, the working group Semantic Web investigates methods from such areas as databases, information retrieval, artificial intelligence and from the Semantic Web community in particular. Germane to this research are applications in challenging fields like eGovernment, eScience, or Big Semantic Data.

People

Daniel Janke

PD Dr. Matthias Thimm

Martin Leinberger

Prof. Dr. Steffen Staab

Software and Services

On one hand, software components have to process huge amounts of data. These data are often weakly structured, get published ad hoc and change frequently. In order to develop software for these data, the developer has to understand the structure of the data and exploit this structure in the software design.

In order to develop software for these data, the developer has to understand the structure of the data and exploit this structure in the software design. However, existing programming environments do not support the software developer in understanding the data and in transferring the data schema into the programming language environment to enable a structured and typed organization of data that helps to avoid run-time errors. On the other hand a large number of software components have to interact with each other in the Web. The adaptation of a soft- ware component, e.g., caused by a change of the data schema, often requires further adaptations of other dependent or interac- ting software components. At this point, the software developer needs novel methods for an automated evolution and coevolution of software components. To tackle these problems, the working group Software and Services in the Web investigates the exploitation of Semantic Web methods to deal with changing data and software components in the development, programming and maintenance of software in order to increase the flexibility and quality of the resulting software.

People

Martin Leinberger

Prof. Dr. Steffen Staab

Web Science und Management

The Web changes how businesses function, cooperate and present themselves. To explain and predict how enterprises (co)operate on the Web one must understand the underlying social processes.

The working group Web Science and Management investigates these social processes on the Web. Its current focus is online reputation of ventures: More and more providers make use of different types of feedback systems by which users can rate the quality of a corporation’s services or products. This accounts for hotel or restaurant booking platforms, online trading and employers. These ratings have become a unique form of currency on the web and have great impact on the reputation of the providers. The working group Web Science and Management investigates die impact of online reputation on the overall reputation of corporations as well as the interaction of employees’ actions in social media and business reputation.

People

JProf. Dr. Mario Schaarschmidt

Web Search and Data Mining

Nowadays, the World Wide Web is both: A huge network of interlinked information as well as a virtual space for people to interact, meet, and share experiences. In the working group on Web Search and Data Mining we use a joint toolset of methods and approaches to analyse both of these aspects. When it comes to information on the Web, a core task is the detection of topics covered by online media. We use probabilistic models to describe topics and to determine the degree of how much a certain online article addresses a topic. For this we do not only analyse the words appearing in the documents, but also the context of the documents, e.g. the place or the social setting. This viewpoint on web contents can be used, for instance, to select a reasonable number of articles which provide the best and most representative coverage of a given topic. In social networks users generate and upload content, but one may also analyse their social interactions. The network structures can be used to predict, for instance, when two users should get into contact with each other, when a user is about to leave a social network and what would be the benefit of extending the ‘like’ button in a social network platform with a ‘dislike’ button.

People

Prof. Dr. Steffen Staab