Research
- JTSKM: Java Implementation of the Time Series Knowledge Mining (TSKM), a pattern extraction process for symbolic interval data. Includes algorithms for calculating Closed Itemsets (DCIclosed, CHARM), Closed Sequential Pattern (BIDE, PrefixSpan) and Closed Partial Orders (algorithm by G.C. Garriga). For preprocessing of numerical time series data and converting into interval data, you can use the MATLAB-Implementation of Fabian Mörchen.
- Contributions to the Databionic MusicMiner: Visualization of large music collections based on a Self-Organizing Map.
- Contributions to the Databionics ESOM Tools: Clustering, visualization, and classification with (Emergent) Self-Organizing Maps.
- Contributions of audio features related classes to the Value-Series-Plugin for RapidMiner (Yale).
- Audio Feature Toolbox: Contibutions to a framework for calculating over 1.200 different audio features and a ranking algorithm.