Filtern
Erscheinungsjahr
- 2016 (31) (entfernen)
Dokumenttyp
- Dissertation (31) (entfernen)
Schlagworte
- Abwasser (1)
- Ad-hoc-Netz (1)
- Algorithmische Geometrie (1)
- Analytischer Hierarchieprozess (1)
- Archäologie (1)
- Beaconless (1)
- Bedrohte Tiere (1)
- Betriebspädagogik (1)
- Bewertungssystem (1)
- Boden (1)
- Controlling (1)
- Daphnia (1)
- Dialog (1)
- Diskurs (1)
- Dispositiv (1)
- Distributed Algorithm (1)
- Drahtloses Sensorsystem (1)
- Drahtloses vermachtes Netz (1)
- E-Participation (1)
- Ebener Graph (1)
- Enterprise Architecture Framework (1)
- Entscheidungstheorie (1)
- Eurokrise (1)
- Evacuation modeling (1)
- Finanzpolitik (1)
- Fiskalpolitik (1)
- Flow decomposition (1)
- Foucault, Michel (1)
- Führung (1)
- Gamebased (1)
- Gamebasiertheit (1)
- Genealogie (1)
- Geographic routing (1)
- Geometric spanner (1)
- Graph (1)
- Grundschule (1)
- Haushaltspolitik (1)
- IT-Security (1)
- Komplexität / Algorithmus (1)
- Kultur (1)
- LIBS (1)
- Laserinduzierte Plasmaspektroskopie (1)
- Local algorithm (1)
- Macht (1)
- Maifisch (1)
- Management (1)
- Medialität (1)
- Medium (1)
- Messung der Ökoeffizienz (1)
- Metamodel (1)
- Nachbarschaftsgraph (1)
- Netzwerktopologie (1)
- OMW (1)
- Olivenölproduktion (1)
- Optimization (1)
- Planar graphs (1)
- Einbringen (1)
- Quasi unit disk graph (1)
- Rationalität (1)
- Reactive algorithm (1)
- Reference Model (1)
- Regieren (1)
- Rhein (1)
- Risiko (1)
- Routing (1)
- Security Requirements (1)
- Semantic Web (1)
- Semantic Web Data (1)
- Servicepersonal (1)
- Southern Amazonia (1)
- Staatsschuldenkrise (1)
- Tourism (1)
- Tourismus (1)
- Umweltleistungsbewertung (1)
- Unit disk graph (1)
- Verteilter Algorithmus (1)
- Wanderfische (1)
- Wireless sensor network (1)
- Wissen (1)
- Wäschereibranche (1)
- Zooplankton (1)
- Zusammenhängender Graph (1)
- decomposition (1)
- description logic (1)
- ecotoxicology (1)
- energetics (1)
- externe Repräsentationen (1)
- field experiment (1)
- flows over time (1)
- fluid disturbances (1)
- fungicide (1)
- greenhouse gases (1)
- kinematics (1)
- land use change (1)
- modeling (1)
- multikriterielles Bewertungssystem (1)
- olive mill wastewater (1)
- olive oil mill wastewater (1)
- pesticide (1)
- problemhaltige Textaufgaben (1)
- rationality (1)
- reasoning (1)
- social system (1)
- soil water repellency (1)
- soils (1)
- soziales System (1)
- stationäre Pflege (1)
- stream (1)
- swarming (1)
- swimming behaviour (1)
- thermal analysis (1)
- zooplankton (1)
- Öko-Effizienz (1)
- Ökoeffizienz (1)
Institut
- Fachbereich 7 (8)
- Fachbereich 8 (4)
- Institut für Informatik (4)
- Fachbereich 5 (2)
- Institut für Erziehungswissenschaft (2)
- Mathematisches Institut (2)
- Arbeitsbereich Biopsychologie, Klinische Psychologie und Psychotherapie (1)
- Institut für Computervisualistik (1)
- Institut für Integrierte Naturwissenschaften, Abt. Physik (1)
- Institut für Kulturwissenschaft (1)
- Institut für Management (1)
- Institut für Mathematik (1)
- Institut für Musikwissenschaft und Musikpädagogik (1)
- Institut für Sozialwissenschaften (1)
- Institut für Umweltwissenschaften (1)
- Institute for Web Science and Technologies (1)
This thesis presents novel approaches for integrating context information into probabilistic models. Data from social media is typically associated with metadata, which includes context information such as timestamps, geographical coordinates or links to user profiles. Previous studies showed the benefits of using such context information in probabilistic models, e.g.\ improved predictive performance. In practice, probabilistic models which account for context information still play a minor role in data analysis. There are multiple reasons for this. Existing probabilistic models often are complex, the implementation is difficult, implementations are not publicly available, or the parameter estimation is computationally too expensive for large datasets. Additionally, existing models are typically created for a specific type of content and context and lack the flexibility to be applied to other data.
This thesis addresses these problems by introducing a general approach for modelling multiple, arbitrary context variables in probabilistic models and by providing efficient inference schemes and implementations.
In the first half of this thesis, the importance of context and the potential of context information for probabilistic modelling is shown theoretically and in practical examples. In the second half, the example of topic models is employed for introducing a novel approach to context modelling based on document clusters and adjacency relations in the context space. They can cope with areas of sparse observations and These models allow for the first time the efficient, explicit modelling of arbitrary context variables including cyclic and spherical context (such as temporal cycles or geographical coordinates). Using the novel three-level hierarchical multi-Dirichlet process presented in this thesis, the adjacency of ontext clusters can be exploited and multiple contexts can be modelled and weighted at the same time. Efficient inference schemes are derived which yield interpretable model parameters that allow analyse the relation between observations and context.