SAS Institute Japan Co., Ltd. JMP Japan Division Tester Yoshiaki Ono
In this presentation, we outline the geometric interpretation in principal component analysis and multiple correspondence analysis.In the context of French stream data analysis that draws on the flow of Benzekri, principal component analysis, correspondence analysis, and multiple correspondence analysis are statistical methods that project points scattered on a multidimensional geometric space onto a low dimensional space is there.On the other hand, in how-to books for beginners in Japan, principal component analysis is often introduced as a method for composing synthetic variables.Quantification III is famous in Japan as a method of applying the same calculation as multiple correspondence analysis, but quantification III is often introduced as a method to maximize the correlation between individual score and category score.These perspectives are somewhat different from the idea in French-style data analysis.
In the report of principal component analysis and multiple correspondence analysis in JMP, the results of France flow (geometric interpretation) and main stream (interpretation of obtaining synthetic variables and scores) are mixed.It seems necessary to have multiple viewpoints when using these functions of JMP.
In this presentation, we briefly introduce the viewpoint of French current data analysis for principal component analysis and multiple correspondence analysis.
[References] Le Roux and Rouanet (2010) "Multiple Correspondence Analysis", SAGE.
