groupingmethod思维导图

这是一篇关于clustering，groupingmethod的思维导图，有助于帮助您熟悉知识要点，加强记忆。有需要的同学，可以收藏下哟。

clustering,distance methods

• introduction
  • No assumption
  • on the basis of similarities or distances
• similarties for item
  • block distance
  • Euclidean distance
  • Minkowski diatance
  • Mahalanobis diatance
  • canberra metric
  • Czekanowski cofficient
• similarities for variables
  • sample correlation coefficients
    • a+d/p
  • cosine
• Hierachical clustering methods
  • agglomerative
    • 1.start with N clusters
    • 2.compute matrix of distance
    • search the distance matrix for the nearest pair of clusters.merge and label new cluster
    • compute diatance between new cluster and the other clusters.repeat step3
    • if number of clusters equals one,then stop
  • division
  • linkage method
    • single linkage
    • complete linkage
    • average linkage
      • between-group linkage
      • within-groups linkage
    • ward's linkage
    • centroid linkage
    • median linkage
• Nonhierachical clustering method
  • feature
  • k-means