Wir bieten laufend Wiss. Teamprojekte/Laborpraktika/Softwareprojekte mit folgenden Schwerpunkten an:
- Data Knowledge Engineering, Data Mining, Maschinelles Lernen
- Information Retrieval
- Visualisierung und Nutzer-Interfaces
Im Folgenden geben wir eine (unvollständige) Übersicht über aktuelle Projekte. Für Anfragen zu diesen und weiteren Themen stehen die Mitarbeiter der Gruppe gerne zur Verfügung. Für eigene Themenvorschläge, die zu unseren Forschungsschwerpunkten passen, sind wir natürlich jederzeit offen. Am besten uns einfach anschreiben!
Hinweise zur Durchführung des Projektes gibt es hier.
Interactive Visualization of Big Networks under Constraints of Importance (Wissenschaftliches Teamprojekt)
A lot of research involves the evaluation of network or graph structures. Especially when evaluating semantic networks, it is important to be able to get an overview and still be able to assess details. A lot of existing programs are built for small networks only and do not support users in understanding bigger structures. This is for example important, if a company provides a taxonomy for their working area, but this taxonomy is too big to be looked at all at once. Additionally, not all users are interested in the same manner in the content of the taxonomy, since they may work only in a specialized field or area. They still need to be able to get the full picture, but may want to define certain parts of the network as more important that others. These definitions could be seen as constraints, that are introduced not only on a visual but also contextual level. It would be therefore interesting, how these constraints could be defined, what kind of constraints are actually helpful and how they could be visualized in a wholesome system. In this project, such a system should be designed and implemented, which can load different network structures via an interface, visualize these networks in different ways (from an overview to a specific view) and let the user define constraints for these networks. Since these networks (especially in the sense of taxonomies) can change quite frequently, a support for visualizing changes between versions of the same network would be helpful to add.
- Define the requirements for a system in order to solve the main task, also think about the kind of constraints a user would expect to set
- Read up on the state of the art of visualizations techniques, user guidance and storing of ontologies, taxonomies and tree structures
- Find tools that visualize similar structures (trees, ontologies, graphs, etc.), e.g. Freemind, protégé
- Compare the tools to your requirements: What are they doing well? What is bad? What is missing?
- Create a system design based on the requirements, the state of the art and the employment of constraints
- Implement a prototype and test it on various structures
- Evaluate the prototype in a user study
- A prototype for an interactive visualization in big graph strutures
- The possibility to load different structures via an interface into the system
- Definition of contraints by the user and a sensible visualization of those
- Support for visualizing changes made between two versions of the same graph
Betreuer: Marcus Thiel