Getting started with streamMOA

The hardware and bandwidth for this mirror is donated by METANET, the Webhosting and Full Service-Cloud Provider.
If you wish to report a bug, or if you are interested in having us mirror your free-software or open-source project, please feel free to contact us at mirror[@]metanet.ch.

Michael Hahsler

The streamMOA package connects stream to clustering algorithms from MOA (Massive Online Analysis). This vignette shows how to create a data stream, process it with a MOA clusterer, and inspect the resulting micro- and macro-clusters. Java 8 or later is required.

Installation

Install the package from CRAN and load it together with stream:

install.packages(c("stream", "streamMOA"))

library(streamMOA)
library(stream)

Create a data stream

DSD_Gaussians() from stream generates a stream with three clusters and a small amount of noise. Request a finite number of observations with get_points():

stream <- DSD_Gaussians(k = 3, d = 2, noise = 0.05)
get_points(stream, n = 5)
#>          X1        X2 .class
#> 1 0.5458535 0.6891478      1
#> 2 0.5751342 0.3031995      2
#> 3 0.5070179 0.2583102      2
#> 4 0.5436979 0.3234242      2
#> 5 0.6294917 0.2378384      2

Stream observations are generated as requested, so plotting or retrieving points advances the stream. Set info = FALSE when only feature values are needed and cluster labels are not required.

plot(stream, n = 500)

Cluster the stream

DSC_CluStream() creates a MOA-based online clusterer. The m parameter sets the maximum number of micro-clusters, while horizon controls the time window. The optional k argument enables weighted k-means reclustering to produce macro-clusters. Use update() to train the clusterer on incoming observations:

clustering <- DSC_CluStream(m = 50, horizon = 100, k = 3)
update(clustering, stream, n = 500)
clustering
#> CluStream 
#> Class: moa/clusterers/clustream/WithKmeans, DSC_MOA, DSC_Micro, DSC 
#> Number of micro-clusters: 50 
#> Number of macro-clusters: 3

Inspect the result

The micro-clusters summarize recent observations. Macro-clusters group those summaries into a smaller number of clusters. Retrieve their centers or plot them with the stream:

get_microclusters(clustering)
#>            X1         X2
#> 1  0.96016355 0.25227255
#> 2  0.52182990 0.65601731
#> 3  0.56865733 0.27717631
#> 4  0.59170698 0.59359064
#> 5  0.47697525 0.70780938
#> 6  0.56252095 0.65414713
#> 7  0.62859918 0.64503513
#> 8  0.94575395 0.26314770
#> 9  0.90051402 0.32183734
#> 10 0.47266308 0.26457753
#> 11 0.06906004 0.78503328
#> 12 0.49812901 0.73775921
#> 13 0.88587126 0.32857369
#> 14 0.67503717 0.03219842
#> 15 0.66036942 0.33072764
#> 16 0.95921384 0.23785634
#> 17 0.51005490 0.22444148
#> 18 0.85708086 0.39108658
#> 19 0.46590908 0.75736211
#> 20 0.49915306 0.83779142
#> 21 0.91234022 0.31101102
#> 22 0.54019816 0.31004215
#> 23 0.56780590 0.29400499
#> 24 0.94228753 0.27649655
#> 25 0.58861360 0.28917296
#> 26 0.97970058 0.22987894
#> 27 0.54200968 0.63802303
#> 28 0.46254419 0.73116493
#> 29 0.45423587 0.74317222
#> 30 0.51865811 0.68553151
#> 31 0.53498163 0.41515597
#> 32 0.48949698 0.74816497
#> 33 0.56540750 0.32188795
#> 34 0.54953709 0.70341318
#> 35 0.94582852 0.11728255
#> 36 0.48188183 0.65824329
#> 37 0.51111227 0.71930316
#> 38 0.53377445 0.37971801
#> 39 0.38967593 0.97536916
#> 40 0.97494889 0.21642202
#> 41 0.54640408 0.29501257
#> 42 0.92216660 0.27849330
#> 43 0.51054455 0.23977742
#> 44 0.53112220 0.67558870
#> 45 0.76448880 0.42889785
#> 46 0.49323525 0.72460230
#> 47 0.50321580 0.66157570
#> 48 0.25899523 0.01903151
#> 49 0.91534332 0.29725388
#> 50 0.97447349 0.23670116
get_macroclusters(clustering)
#>           X1        X2
#> 1 0.06906004 0.7850333
#> 2 0.79085288 0.2770008
#> 3 0.50634679 0.7078762
plot(clustering, stream, type = "both")

Other clusterers include DSC_DenStream(), DSC_ClusTree(), DSC_DStream_MOA(), DSC_BICO_MOA(), and DSC_StreamKM(). See their help pages and the package vignette Introduction to streamMOA for algorithm details, options, and additional examples.

These binaries (installable software) and packages are in development.
They may not be fully stable and should be used with caution. We make no claims about them.