using MIFAMD in order to do MFA
87 views
Skip to first unread message
annemari...@gmail.com
Sep 28, 2020, 3:46:46 PM9/28/20
to FactoMineR users
Dear all, dear François & Julie,
I have a dataset with both quantitative (actually, they are counts, hence the integers) and categorical variables, on which I want to use MFA. However, there are NA's for both groups of variables, so I want to use MIFAMD to impute the missing datapoints. I add the summary below.
I can't get MIFAMD to work, this is what I have used so far:
MIFAMD_output <- MIFAMD(NC_coding4_sel, ncp = 1, maxiter = 999)
- whatever value I put for nap, I get (with variations):
"Error in Residu[, seq(ncol(Xquanti))] : incorrect number of dimensions"
Can you help me out to get MIFAMD working?
Many thanks!
Annemarie
How_many_SG_NCF How_many_PL_NCF How_many_invariant_NCF How_many_paired_NCF
Min. : 1.00 Min. : 1.000 Min. :1.000 Min. : 1.00
1st Qu.:11.00 1st Qu.: 6.000 1st Qu.:1.000 1st Qu.: 4.00
Median :12.00 Median : 7.000 Median :1.000 Median :13.00
Mean :10.79 Mean : 6.371 Mean :2.058 Mean :10.53
3rd Qu.:13.00 3rd Qu.: 7.000 3rd Qu.:3.000 3rd Qu.:16.00
Max. :14.00 Max. :10.000 Max. :9.000 Max. :17.00
NA's :7 NA's :1
How_many_Sg_Agr_CL How_many_PL_Agr_CL How_many_Invariant_Agr_CL
Min. : 1.00 Min. : 1.000 Min. :1.00
1st Qu.: 9.00 1st Qu.: 6.000 1st Qu.:1.00
Median :10.00 Median : 6.000 Median :1.00
Mean : 8.93 Mean : 6.035 Mean :1.91
3rd Qu.:11.00 3rd Qu.: 7.000 3rd Qu.:3.00
Max. :12.00 Max. :10.000 Max. :5.00
NA's :6 NA's :6 NA's :12
How_many_paired_AGR_CL Gender_on_Attr_Adj Gender_on_Dem Gender_on_Num
Min. : 1.000 : 0 : 0 : 0
1st Qu.: 4.000 NA : 0 NA : 0 NA : 0
Median :11.000 no : 25 no : 15 no : 15
Mean : 9.052 yes :129 yes :156 yes :147
3rd Qu.:13.000 NA's: 24 NA's: 7 NA's: 16
Max. :14.000
NA's :24
Gender_on_Quant Gender_on_Poss Gender_on_Wh_Words Gender_on_V Gender_on_Pred_Adj
: 0 : 0 : 0 : 0 : 0
NA : 0 NA : 0 NA : 0 NA : 0 NA : 0
no : 20 no : 20 no :37 no : 25 no : 23
yes :119 yes :149 yes :94 yes :141 yes : 49
NA's: 39 NA's: 9 NA's:47 NA's: 12 NA's:106
Gender_on_Cop Gender_on_Rel Gender_on_Ind_Pers_Pro Gender_on_Poss_Pro
: 0 : 0 : 0 : 0
NA : 0 NA : 0 NA : 0 NA : 0
no : 25 no :21 no : 35 no : 22
yes : 53 yes :99 yes :114 yes :142
NA's:100 NA's:58 NA's: 29 NA's: 14
Gender_on_Dem_Pro Gender_on_Refl_Pro Gender_on_Other AC_on_Attr_Adj AC_on_Dem
: 0 : 0 : 0 : 0 : 0
NA : 0 NA : 0 NA : 0 NA : 0 NA : 0
no :18 no :67 no :53 no :107 no :107
yes :82 yes :13 yes :26 yes : 12 yes : 10
NA's:78 NA's:98 NA's:99 NA's: 59 NA's: 61
AC_on_Num AC_on_Quant AC_on_Poss AC_on_Wh_words AC_Sbj_Agr AC_Obj_Agr
: 0 : 0 : 0 : 0 : 0 : 0
NA : 0 NA : 0 NA : 0 NA : 0 NA : 0 NA : 0
no :108 no :100 no :107 no :89 no :92 no :96
yes : 7 yes : 5 yes : 7 yes :15 yes :34 yes :18
NA's: 63 NA's: 73 NA's: 64 NA's:74 NA's:52 NA's:64
AC_on_Pred_Adj AC_on_Cop AC_on_Rel AC_on_Pers_Pro AC_on_Poss_Pro AC_on_Dem_Pro
: 0 : 0 : 0 : 0 : 0 : 0
NA : 0 NA : 0 NA : 0 NA : 0 NA : 0 NA : 0
no :76 no :76 no :91 no :93 no :107 no :97
yes : 5 yes : 8 yes : 3 yes :16 yes : 10 yes : 4
NA's:97 NA's:94 NA's:84 NA's:69 NA's: 61 NA's:77
AC_on_Refl_Pro AC_on_Other AC_Obl_NP AC_Obl_Everywhere Only_Number_Agr
: 0 : 0 : 0 : 0 : 0
NA : 0 NA : 0 NA : 0 NA : 0 NA : 0
no :88 no :95 no :97 no :111 no :129
yes : 1 yes : 2 yes :17 yes : 4 yes : 2
NA's:89 NA's:81 NA's:64 NA's: 63 NA's: 47
Only_Number_NCF Animacy_Number_NFC Extra_NCF_Animacy Extra_NCF_Number
: 0 : 0 : 0 : 0
NA : 0 NA : 0 NA : 0 NA : 0
no :130 no :125 no :133 no :130
yes : 3 yes : 8 yes : 2 yes : 3
NA's: 45 NA's: 45 NA's: 43 NA's: 45
Extra_NCF_Animacy.Number
: 0
NA : 0
no :132
yes : 1
NA's: 45
Francois Husson
Oct 3, 2020, 12:10:50 PM10/3/20
to factomin...@googlegroups.com
Dear AnnMarie,
Maybe the problem is due o the fact that some categorical variables have some categories with 0 individuals. So you first have to use the droplevels function to suppress these categories.
NC_coding4_sel <- droplevels(NC_coding4_sel)
FH
Maybe the problem is due o the fact that some categorical variables have some categories with 0 individuals. So you first have to use the droplevels function to suppress these categories.
NC_coding4_sel <- droplevels(NC_coding4_sel)
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/7f867231-3ba8-4bc4-be35-d57cfe975b45n%40googlegroups.com.
--
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
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
annemari...@gmail.com
Oct 5, 2020, 8:40:45 AM10/5/20
to FactoMineR users
Dear François,
Yes, you were right! This actually results, depending on the subset of the dataset, in some categories have no variation, so I got some other error messages but I realised the problem and now all is good. I have output from MIFAMD - but which part of it do I feed into MFA()?
This is my call:
mfa1 <- MFA(NC_coding5_sel, group = c(8, 39), type = c("c","n"), name.group = c("continuous","categorical"), tab.comp = ???, graph = F)
I have tried in the place of ??? MIFAMD_output$res.imputeFAMD$completeObs, MIFAMD_output$res.imputeFAMD, MIFAMD_output$res.MI, and all kinds of variation, without success.
Many thanks for all your help!
With best wishes,
Annemarie
Op zaterdag 3 oktober 2020 om 18:10:50 UTC+2 schreef François Husson:
annemari...@gmail.com
Oct 13, 2020, 8:30:33 AM10/13/20
to FactoMineR users
Please, does anybody have a solution regarding how to give MIFAMD output to MFA? (See above)
Many thanks,
Annemarie
Op maandag 5 oktober 2020 om 14:40:45 UTC+2 schreef annemari...@gmail.com:
Reply all
Reply to author
Forward
0 new messages