Multivariate Analysis

​

Objectives: 

Know the main statistical techniques for the analysis of multidimensional phenomena, identify the optimal technique in relation to the type of data and the objectives of the analysis, interpret the results appropriately.

 

Techniques presented: 

Techniques of multivariate statistical analysis for segmentation (cluster analysis), classification (discriminant analysis, decision trees) and for perceptual mapping (factor analysis, correspondence analysis, multidimensional scaling).

​

Tutorials:

​There are exercises for each of the topics covered.

 

Prerequisites: 

Attending the TSC course or having knowledge of the topics it contains is preparatory.

​

Subjects:

Segmentation models

• The interdependence between the variables

  • The distances

• cluster analysis

  • Hierarchical methods

  • The method of k means

  • Example of application of Cluster Analysis

 

Classification models

• Discriminant analysis

• Introduction to decision trees

  • Algorithms (CHAID, C&RT)

  • Stopping rules

  • Risk estimation

  • Profits, Costs of Misclassification, Priors,
    surrogates

 

Perceptual mapping and scaling models

• The factor analysis

  • Steps in performing factor analysis

  • Application example of Principal Components Analysis

• Correspondence analysis

  • The dimensions and the perceptual map

• The optimal scaling

  • The homogeneity analysis

  • Nonlinear principal component analysis (outline)

  • Nonlinear canonical correlation analysis (outline)

• Scaling  multidimensional

  • Reliability analysis

  • The MDS