Chapter 5 Barycentric Discriminant Analysis

5.1 Description

Barycentric discriminant analysis(BADA) is a robust version of discriminant analysis that is used to assign, top re-defined groups(also called categories), observations described by multiple variables. By contrast with traditional discriminant analysis, BADA can be used even when the number of observations is smaller than the number of variables. This makes BADA particularly suited for the analysis of Big Data.

5.2 Heatmap

5.3 Scree Plot

Gives amount of information explained by corresponding component. Gives an intuition to decide which components best represent data in order to answer the research question.

P.S. The most contribution component may not always be most useful for a given research question.

5.4 Factor Scores

  • With Tolerance Interval

5.5 Loadings

5.6 Most Contributing Variables

  • With Bootstrap Ratio

5.7 Permutation Test

5.8 Parallet Test

5.9 Bootstrap Test

5.10 Conclusion

Methods Unhappy Normal Very Happy Reliability
BADA Temp Rain Rain Components have significant contribution but convex hull has overlapping areas