Generate a CA biplot from a 2x2 table

[ivan-marroquin]

Hi,

Thanks for such great package!

I have a question: can I generate a "pseudo" CA biplot from a 2x2 table? Since the table does not have enough dimensions to generate a biplot, I though about adding one extra column/row with 0.5 value in each cell as follows:

library(factoextra) data("housetasks") library(FactoMineR)

extract a 2x2 table from data set contingency <- housetasks[1:2,1:2]

expand 2x2 table to include a dummy column/row with 0.5 in each cell

contingency[,'dummy_col'] <- 0.5 contingency['dummy_row',] <- 0.5

run CA analysis and generate plot

res.ca <- CA(contingency, graph= TRUE)

Would make sense to do a such modification from the perspective of CA? if so, how would you recommend to skip the dummy column/row when generating the pseudo CA biplot?

Thanks for your help,

Ivan

[husson]

Hi, No it is not interesting to analyse a 2x2 table with CA because the points are just on 1 dimension. FH

Le 16/09/2020 à 16:33, ivan-marroquin a écrit :

Hi,

Thanks for such great package!

I have a question: can I generate a "pseudo" CA biplot from a 2x2 table? Since the table does not have enough dimensions to generate a biplot, I though about adding one extra column/row with 0.5 value in each cell as follows:

library(factoextra) data("housetasks") library(FactoMineR)

extract a 2x2 table from data set contingency <- housetasks[1:2,1:2]

expand 2x2 table to include a dummy column/row with 0.5 in each cell

contingency[,'dummy_col'] <- 0.5 contingency['dummy_row',] <- 0.5

run CA analysis and generate plot

res.ca <- CA(contingency, graph= TRUE)

Would make sense to do a such modification from the perspective of CA? if so, how would you recommend to skip the dummy column/row when generating the pseudo CA biplot?

Thanks for your help,

Ivan

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/husson/FactoMineR/issues/7, or unsubscribe https://github.com/notifications/unsubscribe-auth/AB2XXYDAJOEKTQOD3JJTENDSGDEC3ANCNFSM4RO5WUCQ.

-- Francois Husson Department Statistics & Computer science AGROCAMPUS OUEST 65 rue de St-Brieuc - 35042 RENNES Tel: +33 2 23 48 58 86 https://husson.github.io https://husson.github.io/

[ivan-marroquin]

Hi @husson,

Thanks for your reply. I understand your point; however, it would be still interesting to see how rows and columns are far apart (or even close together). I attached the generated plot using the commands from my original post.

Ivan