Automatic Structure Discovery and Visualization of Ontologies

Institut

Lehrstuhl für Bioseparation Engineering (TUM-ED)

Typ

Bachelorarbeit / Semesterarbeit / Masterarbeit /

Inhalt

theoretisch / konstruktiv /  

Beschreibung

Ontologies represent a domain as a network of concepts and relationships. As these networks grow, their visualizations can become difficult to navigate, making it hard to see how concepts are organized and how they connect.

In this thesis, you will investigate methods for automatically identifying and visualizing groups of related concepts in a bioprocess ontology. Because concepts can belong to several meaningful groups, the work will explore ways to reveal overlapping relationships and different levels of detail. You will compare graph-based methods, build an interactive prototype using an existing ontology viewer, and evaluate how well it supports exploration of the network.

Tasks

• Review limitations of the existing approach and relevant graph-clustering and abstraction methods.
• Compare methods for representing relationship types, overlapping groups, and different levels of detail.
• Develop an interactive prototype for exploring the discovered structure.
• Evaluate semantic usefulness, visual clarity, stability, scalability, and practical usability.

Voraussetzungen

• Student in computer science, data science, mathematics, engineering, or a related field.
• Interest in graph algorithms, knowledge representation, and information visualization.
• Programming experience, ideally in Python.
• Willingness to explore an open research problem independently.
• Previous knowledge of bioprocess engineering is not required.

Möglicher Beginn

sofort

Kontakt

Chiara Turrina
Raum: 2404
Tel.: 15767
c.turrinatum.de

Ausschreibung