統計数理研究所共同利用研究 (一般研究2) 研究集会「質的データ分析への再接近4－基本理解と周辺理論－」

Place: The Institute of Statistical Mathematics, Tokyo, Japan.
Seminar room 2 (3rd floor D304)

Program

10:45-12:15
Lecture 1
"An Introduction to Correspondence Analysis"

13:15-15:15
Lecture 2
"Simple Correspondence Analysis and some of its Variants"

15:30-17:00
Lecture 3
"Beyond the introduction: a brief discussion of Multiple Correspondence Analysis and Multi-way Correspondence Analysis"

17:00-17:30
Discussion

Abstract: This tutorial is an introduction to simple and multiple Correspondence Analysis (CA) to analyse categorical data. After reviewing the main contributions of CA literature, we illustrate some methods of quantification of CA, which can also justify the different names with which CA is known in the world.
A brief overview of the different ways of data visualization and its appropriateness (dimension number) for assessing data association will be given. Further, some intriguing variants of simple CA will be illustrated, discussing the symmetric/asymmetric variable relationship, and highlighting the nominal and/or ordinal variables of the analysis.
Finally, the case of multiple and multi-way frequency tables will be considered to study variable association. In particular you will see Multiple Correspondence Analysis using indicator or Burt table, Joint CA and at last (but not least), three-way CA which allows a partitioning of variable interactions. Different data examples, well known in literature, and results will be presented and discussed.

※ 本研究集会は、統計数理研究所共同利用研究「大規模データの特徴抽出と情報表現の研究」(29-共研-2013, 代表者 : 森裕一(岡山理科大学)) からの援助も受けて開催されます。