K-Means Click on the name of a technique to access the corresponding entry of the Glossary 1 Click on a number to access the detailed Table of Contents of the Tutorial

 DESCRIPTIVE MODELING   Correspondence Analysis Principal Components Analysis Hierarchical clustering K-means

 Tutorials 1 Overview of Correspondence Analysis   Independence and interaction between categorical variables.    The two PCAs on contingency tables.    What is expected from a graphic representation ? 2 Mechanism of Correspondence Analysis    Reformating data.     The Chi-square distance.     The two PCAs on rows and columns. 3 Interpretation of Correspondence Analysis     Plots.     Interpretation of the total inertia.     Eigenvalues.     Inertia of the modalities.     Quality of representation of the modalities.     Inertia of the factors. 4 An example (1)   The data.     Inertia.     Interpretation of the factors.     Summary of interpretation of the factors.
 Tutorials 5 An example (2): interpreting the modalities    Quality of representation, square cosines.     Distances to the origin.     Heavy modalities.     Neighboring modalities. 6 An example (3) : the combined plot        The basic idea.     Neighboring modalities.     Confirming with the contingency table.     Summary of the analysis. 7 Complements     Supplementary variables.    Ordinal variables.     The Guttman effect.

 Tutorials 1 Overview of Principal Components Analysis    Maximizing projected inertia.     PCA on observations and on variables.     Interpretation of a PCA.     Other applications. 2 PCA on observations    Maximizing the projected inertia on an axis.   Principal Components and factorial planes.   Quality of representation and influential observations 3 PCA on variables    The space of variables.    "Distance" between two variables.    Circle of correlations.    Detecting correlated, anticorrelated   and uncorrelated variables.
 Tutorials 4 Interpretion of a PCA    Quality of a Principal Component.    Interpreting a Principal Components.    Plots of observations and plots of variables.    Eigenvalues and communalities. Other applications of PCA    Data compression and reconstruction.    Data pre-processing.

_______________________________________________________

CLUSTERING

 Tutorial 1 ASCENDING HIERARCHICAL CLUSTERING Distances between observations.     Distances between clusters. The Ward distance.         The agglomerative algorithm.     Presentation of the hierarchy.     Selecting a typology.

 Tutorials 1 The Minimum Distance rule.     Sum-of-squares criterion.    The K-means algorighm.    Incremental K-means.
 Tutorial Initialization of K-means (number and positions of centroids).   Interpretation aids.

___________________________________________________________________________________