This project will involve implementing nodes for the KNIME Analytics Platform that would aid evaluating clustering performance, and detecting outliers, in addition to other clustering algorithms.

This project will be split into several substages:

• Implementation of clustering analysis metrics: Silhouette Coefficient and Davies-Bouldin Index. These will be the first nodes to be implemented and I will need some time to get familiar with KNIME Node development, thus this will take a bit more time. Estimated time necessary: 2-3 weeks.

• Implementation of Fast-MCD for outlier detection. Estimated time necessary: 2 weeks.

• Implementation of an interactive interface for analyzing clustering performance. Estimated time necessary: 3 weeks.

• Implementation of at least one more clustering algorithm: K-means– and/or COD. More will be implemented if spare time is left. Estimated time necessary: 3 weeks.

Organization

Student

Rytis Kumpa

Mentors

  • Adrian Nembach
  • Martin Horn
close

2019