DSC_EMM {rEMM}

DSC Interface for EMM and tNN (package stream)

Description

Provides Data Stream Clusterer (DSC) interfaces for EMM and tNN so they can be used in the stream framework.

Usage

DSC_EMM(formula = NULL, threshold = 0.2, measure = "euclidean", distFun = NULL,
    centroids = identical(tolower(measure), "euclidean"),
    lambda = 0)

DSC_tNN(formula = NULL, threshold = 0.2, measure = "euclidean",
    centroids = identical(tolower(measure), "euclidean"), lambda = 0)

get_EMM(dsc)
set_EMM(dsc, x)

Arguments

formula

NULL to use all features in the stream or a model formula of the form ~ X1 + X2 to specify the features used for clustering. Only ., + and - are currently supported in the formula.

threshold

A "numeric" with the dissimilarity threshold used by the clustering algorithm for assigning a new observation to existing clusters.

measure

A "character" containing the name of the dissimilarity measure used (see dist in proxy for available measures).

distFun

Specify a function passed on as method to dist in proxy (see dist in proxy). The character string passed on as measure will be used as the measure's name.

centroids

A "logical" indicating if centroids are used for clusters. If FALSE, pseudo medians (first observation of a cluster) are used to represent a cluster.

lambda

A "numeric" specifying the rate for fading.

dsc

an object of class "DSC_EMM".

x

an object of class "EMM"

Details

DSC_tNN and DSC_EMM wrap the clustering algorithms so they can be used with the stream framework. See DSC for details.

get_EMM() and set_EMM() can be used to access the EMM object inside the DSC_EMM object.

Value

An object of class "DSC_EMM" or "DSC_tNN".

Examples

library(stream)

### tNN clustering example

stream <- DSD_Gaussians()
stream

cl <- DSC_tNN(threshold = .1)
cl

update(cl, stream, 100)
cl

get_centers(cl)
get_weights(cl)

plot(cl, stream)

## EMM clustering example
data("EMMsim")
plot(EMMsim_train, pch = NA)
lines(EMMsim_train, col = "gray")
points(EMMsim_train, pch = EMMsim_sequence_train)

stream <- DSD_Memory(EMMsim_train)
stream

cl <- DSC_EMM(threshold = 0.1, measure = "euclidean", lambda = .1)
update(cl, stream, n = 200)
cl

reset_stream(stream)
plot(cl, stream, n = 200, method = "pca")

# inspect and recluster the EMM in the DSC_EMM object
emm <- get_EMM(cl)
plot(emm)

emm <- recluster_hclust(emm, k = 4, method = "average")
plot(emm)

set_EMM(cl, emm)

reset_stream(stream)
plot(cl, stream, n = 200, method = "pca")