Data Mining Project Ideas with ELKI

You are welcome to help ELKI by contributing algorithms.

Some ideas of what to do can be found here in the following list.

To contribute, fork ELKI on github, write your algorithm there, and send a pull request.

Low difficulty

The following list are project ideas for students that want to get started with developing data mining algorithms in ELKI.

k-means variations

Hierarchical clustering variations

DBSCAN and OPTICS extensions

Other clustering algorithms

Historic clustering algorithms

These probably are primarily of interest to the history of clustering.

Outlier detection algorithms

Association rule mining

Data indexing

Evaluation

Visualization

Infrastructure

Medium difficulty

Variational Bayes:

Algorithms for categorial data:

Master level

These projects will require substantial changes and refactorings to the codebase: