Agglomerative clustering is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Read the full entry →Research Wiki · Methods Lab
Cluster analysis and segmentation techniques
Algorithms that group observations by similarity, followed by stability, interpretability and actionability checks.
Methods family
Algorithms that group observations by similarity, followed by stability, interpretability and actionability checks.
Use methods only after defining the estimand, data structure, assumptions, validation plan and decision consequence.
Divisive clustering is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Read the full entry →Single linkage is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Read the full entry →Complete linkage is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Read the full entry →Average linkage is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Read the full entry →Centroid linkage is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Read the full entry →Median linkage is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Read the full entry →Ward’s method is a hierarchical clustering criterion that merges clusters to minimise the increase in within-cluster variance. Results depend on distance definition, scaling and the cut chosen for the dendrogram.
Read the full entry →K-means partitions observations into a chosen number of clusters by minimising within-cluster squared distances to centroids. It is sensitive to scale, initialisation, outliers and the selected number of clusters.
K-medoids is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
PAM clustering is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
CLARA is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
CLARANS is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Fuzzy C-means is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Possibilistic clustering is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Gaussian-mixture models is a method within cluster analysis and segmentation techniques. Algorithms that group observations by similarity, followed by stability, interpretability and actionability checks. Its usefulness depends on data structure, assumptions, validation and whether the output answers the intended decision.
Latent-class analysis estimates unobserved categorical groups from patterns in observed categorical indicators. The classes are probabilistic and require checks for fit, stability, interpretation and external usefulness.
Read the full entry →Latent-profile analysis is a method within cluster analysis and segmentation techniques. Algorithms that group observations by similarity, followed by stability, interpretability and actionability checks. Its usefulness depends on data structure, assumptions, validation and whether the output answers the intended decision.
Read the full entry →Finite-mixture modelling is a method within cluster analysis and segmentation techniques. Algorithms that group observations by similarity, followed by stability, interpretability and actionability checks. Its usefulness depends on data structure, assumptions, validation and whether the output answers the intended decision.
Bayesian mixture models is a method within cluster analysis and segmentation techniques. Algorithms that group observations by similarity, followed by stability, interpretability and actionability checks. Its usefulness depends on data structure, assumptions, validation and whether the output answers the intended decision.
DBSCAN forms clusters from dense regions separated by sparser areas and can label isolated observations as noise. It can find irregular shapes but is sensitive to neighbourhood and density parameters.
HDBSCAN is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
OPTICS is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Spectral clustering is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Affinity propagation is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Mean-shift clustering is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Community detection is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Louvain clustering is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Leiden clustering is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Two-step clustering is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Self-organising maps is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Neural-network clustering is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Co-clustering is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Biclustering is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Consensus clustering is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Ensemble clustering is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Constrained clustering is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Semi-supervised clustering is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Longitudinal clustering is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Sequence clustering is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Trajectory clustering is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Elbow method is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Silhouette coefficient is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Calinski-Harabasz index is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Davies-Bouldin index is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Gap statistic is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Cluster stability is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Bootstrap validation is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Holdout validation is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Segment profiling is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Segment typing algorithms is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Browse linked method entries
Agglomerative clustering
Agglomerative clustering is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodDivisive clustering
Divisive clustering is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodSingle linkage
Single linkage is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodComplete linkage
Complete linkage is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodAverage linkage
Average linkage is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodCentroid linkage
Centroid linkage is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodMedian linkage
Median linkage is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodWard’s method
Ward’s method is a hierarchical clustering criterion that merges clusters to minimise the increase in within-cluster variance. Results depend on distance definition, scaling and the cut chosen for the dendrogram.
Open methodK-means clustering
K-means partitions observations into a chosen number of clusters by minimising within-cluster squared distances to centroids. It is sensitive to scale, initialisation, outliers and the selected number of clusters.
Open methodK-medoids
K-medoids is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodPAM clustering
PAM clustering is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodCLARA
CLARA is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodCLARANS
CLARANS is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodFuzzy C-means
Fuzzy C-means is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodPossibilistic clustering
Possibilistic clustering is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodGaussian-mixture models
Gaussian-mixture models is a method within cluster analysis and segmentation techniques. Algorithms that group observations by similarity, followed by stability, interpretability and actionability checks. Its usefulness depends on data structure, assumptions, validation and whether the output answers the intended decision.
Open methodLatent-class analysis
Latent-class analysis estimates unobserved categorical groups from patterns in observed categorical indicators. The classes are probabilistic and require checks for fit, stability, interpretation and external usefulness.
Open methodLatent-profile analysis
Latent-profile analysis is a method within cluster analysis and segmentation techniques. Algorithms that group observations by similarity, followed by stability, interpretability and actionability checks. Its usefulness depends on data structure, assumptions, validation and whether the output answers the intended decision.
Open methodFinite-mixture modelling
Finite-mixture modelling is a method within cluster analysis and segmentation techniques. Algorithms that group observations by similarity, followed by stability, interpretability and actionability checks. Its usefulness depends on data structure, assumptions, validation and whether the output answers the intended decision.
Open methodBayesian mixture models
Bayesian mixture models is a method within cluster analysis and segmentation techniques. Algorithms that group observations by similarity, followed by stability, interpretability and actionability checks. Its usefulness depends on data structure, assumptions, validation and whether the output answers the intended decision.
Open methodDBSCAN
DBSCAN forms clusters from dense regions separated by sparser areas and can label isolated observations as noise. It can find irregular shapes but is sensitive to neighbourhood and density parameters.
Open methodHDBSCAN
HDBSCAN is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodOPTICS
OPTICS is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodSpectral clustering
Spectral clustering is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodAffinity propagation
Affinity propagation is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodMean-shift clustering
Mean-shift clustering is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodCommunity detection
Community detection is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodLouvain clustering
Louvain clustering is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodLeiden clustering
Leiden clustering is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodTwo-step clustering
Two-step clustering is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodSelf-organising maps
Self-organising maps is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodNeural-network clustering
Neural-network clustering is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodCo-clustering
Co-clustering is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodBiclustering
Biclustering is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodConsensus clustering
Consensus clustering is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodEnsemble clustering
Ensemble clustering is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodConstrained clustering
Constrained clustering is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodSemi-supervised clustering
Semi-supervised clustering is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodLongitudinal clustering
Longitudinal clustering is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodSequence clustering
Sequence clustering is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodTrajectory clustering
Trajectory clustering is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodElbow method
Elbow method is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodSilhouette coefficient
Silhouette coefficient is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodCalinski-Harabasz index
Calinski-Harabasz index is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodDavies-Bouldin index
Davies-Bouldin index is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodGap statistic
Gap statistic is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodCluster stability
Cluster stability is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodBootstrap validation
Bootstrap validation is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodHoldout validation
Holdout validation is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodSegment profiling
Segment profiling is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodSegment typing algorithms
Segment typing algorithms is a statistical or analytical concept within cluster analysis and segmentation techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodNo entries match this search.