subspaceMOA: Subspace and Projected stream clustering

Description

Description

The subspaceMOA package is an extension of the stream package, focusing on stream clustering for high-dimensional data streams.

To this end, two new data streams are provided: DSD_RandomRBFSubspaceGeneratorEvents, a synthetic data stream for high dimensional data with clusters that only exist in certain subspaces of the data and DSD_SubspaceARFFStream, a utility that reads ARFF files with subspace information.

New subspace stream clustering algorithms for high-dimensional data can be constructed using DSC_ThreeStage in combination with one of DSC_subspaceDenStream and DSC_subspaceCluStream and one of DSC_clique, DSC_p3c, DSC_proclus, DSC_subclu.

Additionally, implementations of the existing subspace stream clustering algorithms DSC_PreDeConStream and DSC_HDDStream are present.

Lastly, functions for interactive visualization with the shiny package are also included: A specific clustering can be plotted using plot_stream_interactive and the clustering can also be animated with the function animate_stream_interactive.


subspaceMOA documentation built on May 29, 2017, 10:50 p.m.