corres.fnc: Correspondence Analysis

Description

Correspondence analysis for a contingency table.

Usage

corres.fnc(xtab)

Arguments

xtab

A data frame cross-tabulating frequencies.

Value

A correspondence object with summary and plot methods. The summary method lists eigenvalue rates and coordinates, correlations and contributions for Factor 1 and Factor 2. By default, only the first six rows of the factor tables are shown. Full tables are obtained by specifying header = FALSE when calling summary. For information on higher dimensions, set the option n to the desired number (e.g., n = 3) within summary. See plot.corres for documentation of plot options.

References

F. Murtagh (2005) Correspondence Analysis and Data Coding with JAVA and R, Chapman & Hall/CRC, Boca Raton, FL.

Examples

Run this code

data(oldFrench)
  oldFrench.ca = corres.fnc(oldFrench)
  oldFrench.ca
  summary(oldFrench.ca, head = TRUE)
  plot(oldFrench.ca)

  # more readable plot
  data(oldFrenchMeta)
  plot(oldFrench.ca, rlabels = oldFrenchMeta$Genre, 
  rcol = as.numeric(oldFrenchMeta$Genre), rcex = 0.5, 
  extreme = 0.1, ccol = "blue")

  # create subset of proze texts

  prose = oldFrench[oldFrenchMeta$Genre=="prose" & 
          !is.na(oldFrenchMeta$Year),]
  proseinfo = oldFrenchMeta[oldFrenchMeta$Genre=="prose" & 
          !is.na(oldFrenchMeta$Year),]
  proseinfo$Period = as.factor(proseinfo$Year <= 1250)

  prose.ca = corres.fnc(prose)
  plot(prose.ca, addcol = FALSE, 
  rcol = as.numeric(proseinfo$Period) + 1, 
  rlabels = proseinfo$Year, rcex = 0.7)

  # and add supplementary data for texts with unknown date of composition
  proseSup = oldFrench[oldFrenchMeta$Genre == "prose" & 
          is.na(oldFrenchMeta$Year),]
  corsup.fnc(prose.ca, bycol = FALSE, supp = proseSup, font = 2, 
  cex = 0.8, labels = substr(rownames(proseSup), 1, 4))

Run the code above in your browser using DataLab

Data engineering and BI courses are free!