This function takes a matrix of cosine similarity values,
where the row and column names are tokens. The values will
be calculated for the columns (so if you have a non-square)
matrix, make sure the columns are the values you want.
Arguments
- similarity
the matrix of cosine values.
- n
the number of most related cosine values you want
to return. Remember that you will get token-token cosine back
as a value of 1, so use n+1 to get the number you want.
Value
A dataframe of token-token-cosine values.
Examples
# af_top_sim <- top_n(af_cosine, 6)