ELKI creates an environment for developing KDD applications supported by index structures, with an emphasis on unsupervised learning. It provides various implementations for cluster analysis and outlier detection. It provides index structures such as R*-tree for performance boosting and scalability. It is widely used in research areas by students and faculties up until now and has been gaining attention from other parties recently.
ELKI uses the AGPLv3 license, and can be found at https://elki-project.github.io/. It is comprised of the following packages:
de.lmu.ifi.dbs.elki.algorithm: Contains various algorithms such as clustering, classification, itemset mining, and so on
de.lmu.ifi.dbs.elki.outlier: Defines an outlier-based algorithm
de.lmu.ifi.dbs.elki.statistics: Defines a statistical analysis algorithm
de.lmu.ifi.dbs.elki.database: This is the ELKI database layer
de.lmu.ifi.dbs.elki.index: This is for index structure implementation
de.lmu.ifi.dbs.elki.data: Defines various data types and database object types