Which approach is the correct one? PCA or MFA?

[Javier Hernando]

Hello users!

My data structure is based in 151 individuals (rows) x 51 variables (columns  = 1 categorical variable made up by 3 categories or groups(OO, NUTS, LFD), and 50 continuous numerical variables). The background of the experiment is based in 3 interventions in which patients go through different treatments. 

My main point is what approach should I take? I've read here a post in which was recommended to take PCA and  the categorical variable should fill the role of the supplementary qualitative variable: 

https://groups.google.com/g/factominer-users/c/hTWoBcF4fdM?pli=1 

On the other hand, as I've seen in other examples, the 'gene' dataset contained in the missMDA package, the MFA approach could be used. So, how is the best approach to deal with my dataset?

If I should take this last approach, I have a problem executing this code, after the imputation process to solve NA issue

res.mfa2 <- MFA(cbind.data.frame(PCA[, 1], res.impute$completeObs), group = c(1, 50), type = c("n", "s"), name.group = c("Intervention", "gene_Expression"), num.group.sup = 1)

Error in apply(Lg[group.actif, ], 2, sum) : dim(X) must have a positive length

Thank you so much!

[Francois Husson]

If you have 2 groups, and with 1 group that has only 1 variable, I think that MFA is not the best method because the groupe with a unique variable will have as many importance than the other group that ha 50 variabes. And moreover, you consider the group with 1 variable as a supplementary group, so there is only 1 group that is active, so you have to perform PCA.

FH

--
Vous recevez ce message, car vous êtes abonné au groupe Google Groupes "FactoMineR users".
Pour vous désabonner de ce groupe et ne plus recevoir d'e-mails le concernant, envoyez un e-mail à l'adresse factominer-use...@googlegroups.com.
Cette discussion peut être lue sur le Web à l'adresse https://groups.google.com/d/msgid/factominer-users/172a588d-4d39-4bcf-aba0-aa48b475a602n%40googlegroups.com.

--

	François Husson Department Statistics & Computer science UMR 6625 IRMAR CNRS 65 rue de Saint-Brieuc, CS 84215, 35042 Rennes Cedex Tel: +33 2 23 48 58 86 https://husson.github.io
En 2022, Agrocampus Ouest devient l'Institut Agro Rennes-Angers.

[Javier Hernando]

Thank you so much for the contribution.