Research Topics

Text Clustering / Text Mining

• I'm currently working on a text mining (text clustering and svm support) module for our KAON Framework using background knowledge in the form of an ontology to enhance the representation of text documents.
• I also investigate how to inspect and analyse the computed cluster. I use the Formal Concept Analysis (FCA) for this. More details about the combination of KMeans and FCA to explain clustering results are described in Explaining Text Clustering Results using Semantic Structures.
• I am doing some evaluation of my clustering approach, which uses background knowledge in the form of WordNet during the preprocessing phase, with:
  • Reuters-21578 Corpus
  • new Reuters Corpus
  • Agrovoc with SVM, cf. the master thesis of Boris Lauser
• In the paper WordNet improves Text Document Clustering we present some new evaluation results using WordNet and the Reuters corpus. There is also the technical report 425 which reports the results in more detail.
• We used the Agrovoc thesaurus and the Agrovoc dataset to find out how good "Automatic multi-label subject indexing in a multilingual environment" will work. If you are interested in the Agrovoc dataset, feel free to send me an e-mail.

Data Mining/Knowledge Discovery

• Application of Data Mining
• Data Mining and Databases
• Clustering in high dimensional space
  • Text Clustering
  • Clustering Telecommuniation Customer on the basis of their calling behaviour (building customer profiles)
• Clustering with background knowledge
• Clustering large databases

Telecommunication

• Analysing customer behaviour in the telecommunication domain
• Customer Segmentation
• Segmentation description (rule learning, decision trees)