4.2. Coding details
From the comments, the same number of individual LTs are built as there are judges.
This codification preserves the information given by each judge on the wines.
Each table has as many columns as different words, called individual words, used in the corresponding individual comments.
Individual words can be homologous from one judge to another, and then correspond to the same global word.
The individual LTs describe the individual wine configurations as provided separately by each judge,
visualized by successive CAs applied to these individual LTs taken one by one.
4.2.1. Building individual LTs. In list format
French panel
Building the separate LT for French judges
The individual LTs describe the individual wine configurations as provided separately by each judge, visualized by successive CAs applied to these individual LTs taken one by one.
The 15 French judges are:
cat(names(baseFr))
FE5 FE6 FE12 FP1 FP3 FP4 FP5 FP6 FP7 FP8 FP9 FP10 FP11 FP12 FP2The 15 TextData analyses for the French judges are then performed by removing the French stopwords and saving them in a list. The same process could have been done using a loop.
res.TD.Fr.list <- lapply(1:ncol(baseFr), function(i) TextData(baseFr,var.text=i, Fmin=1, stop.word.user = str.Fr.stopworduser))
names(res.TD.Fr.list) <- names(baseFr)To do the 15 Correspondence analysis of the “individual tables”:
res.LexCA.Fr.list <- lapply(1:length(res.TD.Fr.list), function(i) LexCA(res.TD.Fr.list[[i]], graph=FALSE))
names(res.LexCA.Fr.list) <- names(baseFr)To print the results of the eigenvalues for each of the 15 French judges:
unlist(lapply(1:length(res.TD.Fr.list), function(i) {
cat("\n\n", "**** French Judge ", i, "- Name: ", colnames(baseFr)[i], "\n" )
cat(summary(res.LexCA.Fr.list[[i]],nword=0,ndoc=0,nsup=0))
} ))
**** French Judge 1 - Name: FE5
Correspondence analysis summary
Eigenvalues
Variance % of var. Cumulative % of var.
dim 1 1.000 30.364 30.364
dim 2 1.000 30.364 60.729
dim 3 0.806 24.483 85.212
dim 4 0.250 7.591 92.803
dim 5 0.237 7.197 100.000
Cramer's V 0.686 Inertia 3.293
**** French Judge 2 - Name: FE6
Correspondence analysis summary
Eigenvalues
Variance % of var. Cumulative % of var.
dim 1 1 25 25
dim 2 1 25 50
dim 3 1 25 75
dim 4 1 25 100
dim 5 0 0 100
Cramer's V 0.756 Inertia 4
**** French Judge 3 - Name: FE12
Correspondence analysis summary
Eigenvalues
Variance % of var. Cumulative % of var.
dim 1 1.000 42.056 42.056
dim 2 0.754 31.697 73.753
dim 3 0.624 26.247 100.000
dim 4 0.000 0.000 100.000
dim 5 0.000 0.000 100.000
Cramer's V 0.583 Inertia 2.378
**** French Judge 4 - Name: FP1
Correspondence analysis summary
Eigenvalues
Variance % of var. Cumulative % of var.
dim 1 1.000 43.386 43.386
dim 2 0.884 38.347 81.733
dim 3 0.421 18.267 100.000
dim 4 0.000 0.000 100.000
dim 5 0.000 0.000 100.000
Cramer's V 0.574 Inertia 2.305
**** French Judge 5 - Name: FP3
Correspondence analysis summary
Eigenvalues
Variance % of var. Cumulative % of var.
dim 1 1.000 38.961 38.961
dim 2 1.000 38.961 77.922
dim 3 0.567 22.078 100.000
dim 4 0.000 0.000 100.000
dim 5 0.000 0.000 100.000
Cramer's V 0.606 Inertia 2.567
**** French Judge 6 - Name: FP4
Correspondence analysis summary
Eigenvalues
Variance % of var. Cumulative % of var.
dim 1 1.000 33.333 33.333
dim 2 1.000 33.333 66.667
dim 3 1.000 33.333 100.000
dim 4 0.000 0.000 100.000
dim 5 0.000 0.000 100.000
Cramer's V 0.655 Inertia 3
**** French Judge 7 - Name: FP5
Correspondence analysis summary
Eigenvalues
Variance % of var. Cumulative % of var.
dim 1 0.615 24.348 24.348
dim 2 0.551 21.813 46.161
dim 3 0.424 16.801 62.962
dim 4 0.327 12.939 75.901
dim 5 0.267 10.570 86.471
Cramer's V 0.601 Inertia 2.525
**** French Judge 8 - Name: FP6
Correspondence analysis summary
Eigenvalues
Variance % of var. Cumulative % of var.
dim 1 0.604 54.374 54.374
dim 2 0.215 19.372 73.746
dim 3 0.146 13.168 86.914
dim 4 0.089 7.995 94.909
dim 5 0.028 2.558 97.467
Cramer's V 0.398 Inertia 1.11
**** French Judge 9 - Name: FP7
Correspondence analysis summary
Eigenvalues
Variance % of var. Cumulative % of var.
dim 1 0.919 23.546 23.546
dim 2 0.825 21.143 44.689
dim 3 0.708 18.153 62.842
dim 4 0.634 16.245 79.087
dim 5 0.407 10.431 89.518
Cramer's V 0.747 Inertia 3.903
**** French Judge 10 - Name: FP8
Correspondence analysis summary
Eigenvalues
Variance % of var. Cumulative % of var.
dim 1 0.760 27.724 27.724
dim 2 0.581 21.197 48.921
dim 3 0.422 15.387 64.308
dim 4 0.364 13.272 77.580
dim 5 0.265 9.659 87.239
Cramer's V 0.626 Inertia 2.74
**** French Judge 11 - Name: FP9
Correspondence analysis summary
Eigenvalues
Variance % of var. Cumulative % of var.
dim 1 0.863 52.025 52.025
dim 2 0.714 43.027 95.052
dim 3 0.082 4.948 100.000
dim 4 0.000 0.000 100.000
dim 5 0.000 0.000 100.000
Cramer's V 0.487 Inertia 1.66
**** French Judge 12 - Name: FP10
Correspondence analysis summary
Eigenvalues
Variance % of var. Cumulative % of var.
dim 1 0.842 43.164 43.164
dim 2 0.627 32.176 75.340
dim 3 0.481 24.660 100.000
dim 4 0.000 0.000 100.000
dim 5 0.000 0.000 100.000
Cramer's V 0.528 Inertia 1.95
**** French Judge 13 - Name: FP11
Correspondence analysis summary
Eigenvalues
Variance % of var. Cumulative % of var.
dim 1 0.780 23.629 23.629
dim 2 0.677 20.504 44.133
dim 3 0.547 16.560 60.693
dim 4 0.511 15.478 76.171
dim 5 0.459 13.887 90.058
Cramer's V 0.687 Inertia 3.303
**** French Judge 14 - Name: FP12
Correspondence analysis summary
Eigenvalues
Variance % of var. Cumulative % of var.
dim 1 1.000 39.604 39.604
dim 2 0.917 36.321 75.925
dim 3 0.608 24.075 100.000
dim 4 0.000 0.000 100.000
dim 5 0.000 0.000 100.000
Cramer's V 0.601 Inertia 2.525
**** French Judge 15 - Name: FP2
Correspondence analysis summary
Eigenvalues
Variance % of var. Cumulative % of var.
dim 1 1 50 50
dim 2 1 50 100
dim 3 0 0 100
dim 4 0 0 100
dim 5 0 0 100
Cramer's V 0.632 Inertia 2 To count the number of French judges for whom the first factorial plane explains more than 70% in order to obtain some information on the inertias of the individual tables.
cat(sum(sapply(1:length(res.TD.Fr.list), function(x) res.LexCA.Fr.list[[x]]$eig[2,3]>70)))
8
Catalan panel
Building the separate LT for Catalan judges The 9 Catalan judges are names(baseCat)
cat(names(baseCat))
CE1 CE2 CE3 CE4 CE7 CE8 CE9 CE10 CE11The 9 TextData analyses for the Catalan judges are then performed by removing the Catalan stopwords and saving them in a list. The same process could have been done using a loop.
res.TD.Cat.list <- lapply(1:ncol(baseCat), function(i) TextData(baseCat,var.text=i, Fmin=1, stop.word.user = str.Cat.stopworduser))
names(res.TD.Cat.list) <- names(baseCat)
To do the 9 Correspondence analysis of the "individual tables":
res.LexCA.Cat.list <- lapply(1:length(res.TD.Cat.list), function(i) LexCA(res.TD.Cat.list[[i]], graph=FALSE))
names(res.LexCA.Cat.list) <- names(baseCat)
To print the results of the eigenvalues for each of the 9 Catalan judges:
unlist(lapply(1:length(res.TD.Cat.list), function(i) {
cat("\n\n", "**** Catalan Judge ", i, "- Name: ", colnames(baseCat)[i], "\n" )
cat(summary(res.LexCA.Cat.list[[i]],nword=0,ndoc=0,nsup=0))
} ))
**** Catalan Judge 1 - Name: CE1
Correspondence analysis summary
Eigenvalues
Variance % of var. Cumulative % of var.
dim 1 1.000 40.000 40.000
dim 2 0.767 30.667 70.667
dim 3 0.733 29.333 100.000
dim 4 0.000 0.000 100.000
dim 5 0.000 0.000 100.000
Cramer's V 0.598 Inertia 2.5
**** Catalan Judge 2 - Name: CE2
Correspondence analysis summary
Eigenvalues
Variance % of var. Cumulative % of var.
dim 1 1.000 33.333 33.333
dim 2 1.000 33.333 66.667
dim 3 1.000 33.333 100.000
dim 4 0.000 0.000 100.000
dim 5 0.000 0.000 100.000
Cramer's V 0.655 Inertia 3
**** Catalan Judge 3 - Name: CE3
Correspondence analysis summary
Eigenvalues
Variance % of var. Cumulative % of var.
dim 1 1.000 37.500 37.500
dim 2 1.000 37.500 75.000
dim 3 0.667 25.000 100.000
dim 4 0.000 0.000 100.000
dim 5 0.000 0.000 100.000
Cramer's V 0.617 Inertia 2.667
**** Catalan Judge 4 - Name: CE4
Correspondence analysis summary
Eigenvalues
Variance % of var. Cumulative % of var.
dim 1 0.897 52.677 52.677
dim 2 0.667 39.158 91.835
dim 3 0.139 8.165 100.000
dim 4 0.000 0.000 100.000
dim 5 0.000 0.000 100.000
Cramer's V 0.493 Inertia 1.702
**** Catalan Judge 5 - Name: CE7
Correspondence analysis summary
Eigenvalues
Variance % of var. Cumulative % of var.
dim 1 1.000 28.504 28.504
dim 2 1.000 28.504 57.007
dim 3 0.856 24.408 81.415
dim 4 0.652 18.585 100.000
dim 5 0.000 0.000 100.000
Cramer's V 0.708 Inertia 3.508
**** Catalan Judge 6 - Name: CE8
Correspondence analysis summary
Eigenvalues
Variance % of var. Cumulative % of var.
dim 1 0.838 38.560 38.560
dim 2 0.750 34.497 73.057
dim 3 0.586 26.943 100.000
dim 4 0.000 0.000 100.000
dim 5 0.000 0.000 100.000
Cramer's V 0.557 Inertia 2.174
**** Catalan Judge 7 - Name: CE9
Correspondence analysis summary
Eigenvalues
Variance % of var. Cumulative % of var.
dim 1 0.719 38.011 38.011
dim 2 0.622 32.883 70.894
dim 3 0.396 20.933 91.827
dim 4 0.155 8.173 100.000
dim 5 0.000 0.000 100.000
Cramer's V 0.52 Inertia 1.893
**** Catalan Judge 8 - Name: CE10
Correspondence analysis summary
Eigenvalues
Variance % of var. Cumulative % of var.
dim 1 1.000 33.333 33.333
dim 2 1.000 33.333 66.667
dim 3 1.000 33.333 100.000
dim 4 0.000 0.000 100.000
dim 5 0.000 0.000 100.000
Cramer's V 0.655 Inertia 3
**** Catalan Judge 9 - Name: CE11
Correspondence analysis summary
Eigenvalues
Variance % of var. Cumulative % of var.
dim 1 1.000 38.298 38.298
dim 2 1.000 38.298 76.596
dim 3 0.611 23.404 100.000
dim 4 0.000 0.000 100.000
dim 5 0.000 0.000 100.000
Cramer's V 0.611 Inertia 2.611
To count the number of Catalan judges for whom the first factorial plane explains more than 70% in order to obtain some information on the inertias of the individual tables:
cat(sum(sapply(1:length(res.TD.Cat.list), function(x) res.LexCA.Cat.list[[x]]$eig[2,3]>70)))
6