Term frequency matrix

Question

I have a string like this:

m<-"abcdabcdbcadacbddabcc..."

I would like to generate a matrix like this:

How can I do that in r?

Answer 1

This gives what I believe you're after:

m <- "abcdabcdbcadacbddabcc"

library(qdap)

chars <- unique(unlist(strsplit(m, "")))
terms <- paste2(expand.grid(rep(list(chars), 3)), sep="")
t(counts(termco(m, match.list=sort(terms)))[, -c(1:2)])

Output:

    1
aaa 0
aab 0
aac 0
aad 0
aba 0
.
.
.
dcc 0
dcd 0
dda 1
ddb 0
ddc 0
ddd 0

Answer 2

The function gregexpr gives you the position of each match of the pattern.

You can do this:

a <- c("a","b","c")
b <- matrix(outer(a,a,paste,sep=""),ncol=1)
patterns <- matrix(outer(a,b,paste,sep=""),ncol=1)

m<-"abcdabcdbcadacbddabcc..."

positions <- function(pattern, text) 
  gregexpr(pattern, text)[[1]][1]

sapply(patterns, positions, text=m)