We review existing models for modelling links between documents and show parallels to the multi-Dirichlet based topic model. The second part of my talk is about applications of multi-Dirichlet models in modelling documents with citations, log files and user profiles of a social network.
AbstractSwarm allows graphical modeling and simulation of problems from different domains as multi-agent systems. Simulations and their results are visualized across problem domains without the need to define a specific scheme for a concrete problem.
We introduce our implementation of the STAR algorithm and recent experiment results of its performance. The STAR algorithm is used for relationship query over large relationship graphs, which can be considered from algorithmic point of view as a Steiner Tree problem.
The Linked Open Data cloud provides a wide range of different types of information which are interlinked and connected. When a user or application is interested in specific types of information under time constraints it is best to ex- plore this vast knowledge network in a focused and directed way.
Motivating contributors is an important task for owners of online communities. Literature provides multiple suggestions to motivate users to increase content contribution.
Structured open data available on the web not only encode heterogamous structural relationship between entities of various types but are also a source of textual information associated with these entities. Governments at various levels can benefit from this historical data while making certain policy decisions. Analysis of text rich structured data require development of hybrid tools that are not only able to exploit the structure of the data but also at the same time make use of the textual and structural information for appropriate segmentation/disaggregation and identification of open or linked open datasets that satisfy specific human information needs in various scenarios.
To use RDF in a distributed fashion, data must be partitioned in an intelligent way to keep communication between storage nodes low and reduce query response times. We investigate the utilization of SPARQL query log analysis as an approximation for possible partitioning approaches.
A graph database for sensor data receives continuous data streams from static or moving sensors. There exist many different sensor types which produce a large variety of different data measured at specific points in time. If all these data is combined, valuable information could be gained. During the talk, a conceptual implementation of a graph database for spatio-temporal sensor data is presented. This implementation is designed to support insertions and spatio-temporal queries efficiently.
Monitoring of server infrastructures is a key aspect of maintaining a distributed architecture. In the talk I will explain the structure and characteristics of telemetry data that is captured by state of the art monitoring systems, and present some approaches for detecting changes or anomalies.
Im Rahmen des Projekt- und Forschungspraktikums "Mobile Sensors" wurde ein System zur Analyse des Radverkehrsflusses entwickelt, indem die, mit handelsüblichen Smartphones, während einer Fahrt gesammelten Sensordaten an einen Server gesendet und dort ausgewertet werden.