3mta3

Hmmm

Topics related to conditional probability: conditional probability, conditioning (probability), conditional probability distribution, conditional expectation, regular conditional probability

Clustering

• cluster analysis
  • clustering high-dimensional data
  • cluster analysis (in marketing)
  • Determining the number of clusters in a data set

Related concepts

• Dimension reduction
• Unsupervised learning
• Density estimation

Main types

• k-means clustering
  • Lloyd's algorithm
  • Linde-Buzo-Gray algorithm
  • k-medoids
• hierarchical clustering
  • Single-linkage clustering
  • UPGMA
• mixture model
  • Chinese restaurant process
  • Dirichlet process
  • Pitman–Yor process

Applications

• Segmentation (image processing)
  • Thresholding (image processing)
    • Otsu's method
• cluster analysis (in marketing) (above)
  • Market segment
• Cluster sampling
• Microtargeting
• Document clustering
  • Latent Dirichlet allocation
• Psychometrics
  • Pathfinder Networks

Other

• Biclustering
• Canopy clustering algorithm (not really an algorithm, more a heuristic for speeding up)
• Constrained clustering (requires certain pairs either must or must not be in same cluster)
• Conceptual clustering
  • Cobweb (clustering)
• Formal concept analysis
• Fuzzy clustering
  • Fuzzy c-means on cluster analysis
• Spectral clustering on cluster analysis
• Locality sensitive hashing
• Artificial neural network (mention, but no explanation)

Comparison

• Silhouette (clustering) (graphical representation)
• Cophenetic correlation
• Rand index
  • Adjusted rand index
• Mutual information
  • Adjusted Mutual Information

Unknown

• Consensus clustering (doesn't really explain what it is)
• Information bottleneck method
• Correlation clustering

Algorithms

N for novel (i.e. vanity page for CS paper), N? for possibly novel

• Ants sleeping model N
• birch (data clustering) N
• Cure data clustering N
• Cluster-weighted modeling N?
• DBSCAN
• FLAME clustering N
• Girvan-Newman algorithm
• k-means++ N? (for choosing starting points for k-means)
• Neighbourhood components analysis N
• OPTICS algorithm N?
• QT (quality threshold) clustering on cluster analysis
• SUBCLU N?