AbstractClustering

Supertype for clusterings.

AbstractClusteringDifference

Supertype for clustering differences.

HierarchicalClustering{Tc<:Integer,Tw<:Real} <: AbstractClustering

Hierarchical clustering.

PartitionalClustering{Tc<:Integer,Tw<:Real,Ty<:Real,Tm<:Real} <: AbstractClustering

Partitional clustering.

PartitionalClusteringDifference <: AbstractClusteringDifference

Difference between two partitional clusterings.

axes(a::AbstractClusteringDifference[, d])

Access the feature and instance identifiers differences. Optionally, specify dimension d to get the identifier difference of that dimension only.

axes(a::AbstractClustering[, d])

Access the feature and instance identifiers. Optionally, specify dimension d to get the identifiers of that dimension only.

instances(a::AbstractClusteringDifference)

Access the instance identifiers difference.

instances(a::AbstractClustering)

Access the instance identifiers.

assignments(a::PartitionalClusteringDifference)

Access the assignments difference.

assignments(a::PartitionalClustering)

Access the assignments.

backwarddiff(a::AbstractVector{<:AbstractClustering}, i::Int[, h::Int=1])

Compute the backward difference of the clustering at index i with step size h.

constraints(a::AbstractClusteringDifference)

Access the contraints difference.

constraints(a::AbstractClustering)

Access the contraints.

features(a::AbstractClusteringDifference)

Access the feature identifiers difference.

features(a::AbstractClustering)

Access the feature identifiers.

forwarddiff(a::AbstractVector{<:AbstractClustering}, i::Int[, h::Int=1])

Compute the forward difference of the clustering at index i with step size h.

kmeans(X::AbstractMatrix{<:Real}, r::AbstractVector{Int},
    c::AbstractVector{Int}, μ::AbstractMatrix{<:Real}; <keyword arguments>)

Cluster the data X with the $k$-means algorithm.

Keyword arguments

• maxiter::Int=256: the number of maximum iterations.
• dist::SemiMetric=SqEuclidean(): the distance function.
• ϵ::AbstractFloat=1.0e-6: the absolute tolerance for convergence.

kmeans(X::AbstractMatrix{<:Real}, μ::AbstractMatrix{<:Real}; <keyword arguments>)

Like kmeans, but automatically generate r and c according to the size of X.

parameters(c::AbstractClusteringDifference)

Access the parameters difference.

parameters(c::PartitionalClustering)

Access the parameters.

weights(a::AbstractClusteringDifference)

Access the weights difference.

weights(a::AbstractClustering)

Access the weights.