Table of Interactions - Case with pets and houses

问题

I have a list of houses and a list of animal species.

houses = c(1,1,2,3,4,4,4,4,5,6,5)
animals = c(\'cat\',\'dog\',\'cat\',\'dog\',\'rat\', \'cat\', \'spider\', \'snake\', \'cat\', \'cat\', \'rat\')

I am trying to create a function that returns an upper triangular table that indicates for each pet, the number of times that it was observed to live in the same house than the other animal species. Does it make sense?

For the above example, the table should look like this (hope there\'s no mistake!) :

    dog   rat   spider   snake
cat  1     2      1        1      
dog        0      0        0
rat               1        1
spider                     1

---

Note: This function should work for any two vectors of same lengths, whatever if they contain numbers or string

回答1:

Use table and crossprod:

out <- crossprod(table(houses, animals))
out[lower.tri(out, diag=TRUE)] <- NA
out
#         animals
# animals  cat dog rat snake spider
#   cat     NA   1   2     1      1
#   dog     NA  NA   0     0      0
#   rat     NA  NA  NA     1      1
#   snake   NA  NA  NA    NA      1
#   spider  NA  NA  NA    NA     NA

Since the output is a matrix you can suppress the printing of the NA values directly in print:

print(out,na.print="")
#         animals
# animals  cat dog rat snake spider
#   cat          1   2     1      1
#   dog              0     0      0
#   rat                    1      1
#   snake                         1
#   spider                         

来源：https://stackoverflow.com/questions/19891278/table-of-interactions-case-with-pets-and-houses