Correspondence Analysis in R

Correspondence analysis (CA) is an extension of principal component analysis (principal-component-analysis) suited to explore relationships among qualitative variables (or categorical data). Like principal component analysis, it provides a solution for summarizing and visualizing data set in two-dimension plots. #the codes # Correspondence Analysis in R library("FactoMineR") library("factoextra") library("gplots") #Data format head(housetasks) #Chi-square test to evaluate row and column chisq =chisq.test(housetasks) chisq #compute correspondence analysis res.ca=CA(housetasks, graph = FALSE) # to get eigenvalue EV=get_eigenvalue(res.ca) EV # to look scree plot fviz_screeplot(res.ca, addlabels = TRUE, ylim = c(0, 50)) # Row variables row= get_ca_row(res.ca) # Coordinates head(row$coord) # Cos2: quality on the factore map head(row$cos2) # Contributions to the principal components head(row$contrib) ##Biplot fviz_ca_row(res.ca, repel = TRUE)# to look only Row fviz_ca_biplot(res.ca, repel = TRUE) #draws a standard asymetric biplot: fviz_ca_biplot(res.ca, map ="rowprincipal", arrow = c(TRUE, TRUE), repel = TRUE) ########################################################### #method-2 -with function ca library(ca) ca1 = ca(housetasks) # sqrt of eigenvalues ca1$sv # row coordinates head(ca1$rowcoord) # column coordinates head(ca1$colcoord) # plot plot(ca1) ##################################################### # metod -3 # CA with function dudi.coa library(ade4) # apply ca ca3 = dudi.coa(housetasks, nf = 5, scannf = FALSE) # sqrt of eigenvalues ca3$eig # row coordinates head(ca3$li) # column coordinates head(ca3$co) ################################# # mthod -4 # PCA with function afc library(amap) # apply CA ca4 = afc(housetasks) # eigenvalues ca4$eig # row coordinates head(ca4$scores) # column coordinates head(ca4$loadings) # plot plot(ca4) ##################################### # method -5 # CA with function corresp library(MASS) # apply CA ca5 = corresp(housetasks,4) # sqrt of eigenvalues ca5$cor # row coordinates head(ca5$rscore) # column coordinates head(ca5$cscore) plot(ca5) ##################################