0

I have encountered a problem with the MFA in FactoMineR. I am working with a data set containing physical, chemical and microbiological continuous variables measured in tomato plants, taken from 2 different treatments and at 3 time points. I have accommodated my data like this:

structure(list(row.names = structure(c(1L, 4L, 7L, 10L, 13L, 
16L), .Label = c("GBA1", "GBA2", "GBA3", "GBB1", "GBB2", "GBB3", 
"GBC1", "GBC2", "GBC3", "GBD1", "GBD2", "GBD3", "GBE1", "GBE2", 
"GBE3", "RWA1", "RWA2", "RWA3", "RWB1", "RWB2", "RWB3", "RWC1", 
"RWC2", "RWC3", "RWD1", "RWD2", "RWD3", "RWE1", "RWE2", "RWE3", 
"RWF1", "RWF2", "RWF3", "RWG1", "RWG2", "RWG3", "RWH1", "RWH2", 
"RWH3", "RWI1", "RWI2", "RWI3", "RWJ1", "RWJ2", "RWJ3"), class = "factor"), 
    Trt = structure(c(2L, 2L, 2L, 2L, 2L, 1L), .Label = c("Mineral", 
    "Organic"), class = "factor"), Status = structure(c(1L, 1L, 
    1L, 1L, 1L, 1L), .Label = c("H", "S"), class = "factor"), 
    Humidity = c(87.21704394, 80.29885961, 65.68047337, 85.9775641, 
    83.33333333, 85.98568282), pH = c(5.44, 5.94, 6.64, 6.19, 
    6.13, 5.45), Conductivity = c(837L, 867L, 752L, 871L, 699L, 
    406L), Nit.N = c(436.18, 433.92, 418.1, 458.78, 411.32, 167.24
    ), Ammonia.N = c(3.8122, 2.6452, 1.945, 1.7116, 2.4896, 7.16
    ), P = c(30.95, 15.2, 20.15, 16.1, 18.35, 48.2), K = c(135, 
    35, 95, 40, 145, 275), Ca = c(1287.5, 1427.5, 1610, 1570, 
    1640, 130), Mg = c(367.5, 575, 537.5, 532.5, 590, 42.5), 
    S = c(705L, 924L, 603L, 962L, 626L, 111L), Sodium = c(92.5, 
    170, 135, 127.5, 137.5, 35), Chlorides = c(15.1, 11.1, 15.4, 
    13.2, 13.8, 10.8), Fe = c(1.5, 2.2, 1.7, 2, 2.1, 3.1), Mn = c(1.1, 
    0.55, 0.7, 0.4, 0.65, 1.9), Rhizobium = c(0, 0, 0, 0, 0, 
    0), Total.bacteria = c(7207207.207, 5454545.455, 22727272.73, 
    18918918.92, 30630630.63, 64864864.86)), .Names = c("row.names", 
"Trt", "Status", "Humidity", "pH", "Conductivity", "Nit.N", "Ammonia.N", 
"P", "K", "Ca", "Mg", "S", "Sodium", "Chlorides", "Fe", "Mn", 
"Rhizobium", "Total.bacteria"), row.names = c(NA, 6L), class = "data.frame")

I divided the variables in categorical (first 2), then the other 16 are continuous. However, I want to treat the 2 categorical variables separately. So I wrote the following code:

>res <- MFA(Oliver, group=c(1,1,3,11,2), type=c("n", "n","s", "s","s"),ncp=5,name.group=c("Sub","Stat", "Phys", "Chem", "Microbial"))

However, it doesn't seem to work. Hence, I tried the following:

>res=MFA(Oliver,group=c(2,16),type=c(rep("n",1),rep("s",1)),ncp=5,name.group=c("cat","cont"))

and this other:

>res=MFA(Oliver,group=c(2, 3, 11,2),type=c(rep("n",1),rep("s",3)), ncp=5,name.group=c("type","Phys", "Chem", "Microbial"))

But I kept having the same problem ("not convenient group definition"). Is there anything that I can do to keep the first 2 categorical groups separately? I would really appreciate your advice on how to properly run the model!

Best wishes,

Emma

Roman Luštrik
  • 69,533
  • 24
  • 154
  • 197
Emma Hernandez Sanabria
  • 9
  • 5
  • I could not reproduce your error with your example dataset. Could you post the full dataset here? Kind regards, FM – FM Kerckhof Oct 16 '13 at 13:52

1 Answers1

0

I think that the problem comes from your variable Status which is not a variable since all the values are equal to "H". So no analysis cn be done with this "variable". You can suppress it, there is no information in this column. And then, it should work.

Francois

Husson
  • 141
  • 3