Fig_09

pdf("fig_09.pdf", width = 8, height = 5)
par(bty = "n", las = 1, mar = c(3, 6, 1, 1) + 0.2, mgp = c(5, 1, 0))

# función suma armónicos
sum_arm <- function(n) {
  a <- numeric(n)
  for (i in 1:n) {
    for (j in 1:i) {
      a[j] <- a[j] + 1/i
    }
  }
  a/n
}

# función para contar la ocurrencia de una letra
cuentaletras <- function(letra, texto) {
  texto2 <- gsub(letra, "", texto, ignore.case = TRUE)
  return (nchar(texto) - nchar(texto2))
}
# cargamos el libro en una cadena gigante
libro <- readChar("quijote.txt", file.info("quijote.txt")$size)
# quitamos saltos de línea
libro <- gsub("\n", "", libro)
# quitamos las tildes y diéresis
libro <- chartr("áéíóúàèìòùäëïöüÁÉÍÓÚÀÈÌÒÙÄËÏÖÜ", "aeiouaeiouaeiouAEIOUAEIOUAEIOU", libro)
# abecedario
abecedario <- "abcdefghijklmnñopqrstuvwxyz"
# contaje y resultado
for (i in 1:27) {
  letra <- substr(abecedario, i, i)
  cuenta <- cuentaletras(letra, libro)
  cat(letra, ": ", cuenta, "\n")
}

# datos
obs <- c(229147, 200466, 162480, 125701, 108415, 100932, 90055, 89125, 87230, 79550, 61734, 59425,  44649, 35451, 32478, 25107, 24142, 19916, 17855, 17224, 10530, 7581, 6490, 4241, 377, 2)
nomb <- c("e", "a", "o", "s", "n", "r", "i", "l", "d", "u", "t", "c", "m", "p", "q", "y", "b", "h", "v", "g", "j", "f", "z", "ñ", "x", "w")
n <- length(obs)
teor <- sum(obs)*sum_arm(n)

R2 <- 1 - sum((obs - teor)^2) / var(obs) / (n - 1)

# gráfico
plot(1:n - 0.1, obs, ylab = "Número de veces", type = "h", las = 2, ylim = c(0, 250000), lwd = 2, xlab = "", xaxt = "n")
axis(1, at = 1:n, labels = nomb, las = 1, cex.axis = 0.8)
lines(1:n + 0.1, teor, type = "h", col = "red", lwd = 2)
legend("topright", bty = "n", cex = 0.8, c("observado", "teórico"), lty = 1, lwd = 2, col = c("black", "red"))
text(16, 160000, bquote(R^2 ~ ":" ~ .(round(R2, 2))))

dev.off()