Data Mining Algorithms in ELKI
The following data-mining algorithms are included in the ELKI 0.8.0 release. For literature references, click on the individual algorithms or the references overview in the JavaDoc documentation. See also RelatedPublications
Clustering Algorithms:
- Affinity Propagation
- CFSFDP Density-Peak Clustering
- Canopy Pre-Clustering
- Leader Clustering
- Density-based clustering:
- Naive Mean-Shift Clustering
- Gaussian Mixture Modeling with EM
- Hierarchical clustering:
- AGNES
- Anderberg
- CLINK
- HACAM: Hierarchical Clustering Around Medoids
- HDBSCAN* with linear memory consumption
- Linear Memory NN-Chain
- Medoid Linkage Clustering
- MiniMax Clustering
- MiniMax with Anderberg acceleration
- MiniMax with NN-Chain
- NN-Chain
- SLINK
- SLINK-based HDBSCAN* with linear memory consumption
- BIRCH Leaf Clustering
- Cluster extraction:
- k-means family:
- Annulus
- Best Of Multiple k-means
- Bisecting
- CompareMeans
- Elkan
- Exponion
- FuzzyCMeans
- GMeans
- Hartigan and Wong
- Hamerly
- Lloyd k-means
- Lloyd k-medians
- MacQueen
- k-d-tree Filtering
- k-d-tree Pruning
- KMeans--
- Shallot
- Simplified Elkan
- SortMeans
- Yin-Yang
- Parallel Lloyd
- BIRCH k-Means
- BETULA k-Means
- Single assignment k-means
- X-Means
- Spherical k-means
- K-Medoids
- Silhouette clustering:
- SNNClustering
- Support Vector Clustering
- Correlation clustering algorithms:
- Subspace (axis-parallel) clustering algorithms:
- Biclustering algorithms:
- Clustering algorithms for 1-dimensional data only:
- Trivial clustering algorithms (for reference and evaluation):
- Uncertain clustering algorithms:
Outlier Detection
- Distance-based outlier detection:
- Density based:
- LOF family of methods:
- Angle-based outlier detection:
- Intrinsic dimensionality based:
- Clustering based outlier detection:
- COP
- DWOF
- GaussianModel
- GaussianUniformMixture
- OPTICSOF
- SimpleCOP
- Support Vector Machine based:
- Subspace outlier detection:
- Spatial outlier detection:
- Meta outlier methods:
- Trivial outlier methods (for reference and evaluation):
Classification algorithms (for use with ClassifierHoldoutEvaluationTask)
Frequent Itemset Mining:
Time Series Analysis:
Projection:
- BarnesHutTSNE
- SNE
- TSNE
- itSNE, via the TSNE class
Other:
Benchmarking algorithms:
Data set statistics: