Sample dataframe included in package. Word pairs are arrayed in columns. Columns need not be immediately adjacent within your dataframe.
| word1 | word2 |
|---|---|
| Dog | trumpet |
| the | BANANA |
| rat | astronaut |
| *&^% | lizard |
| bird | bird |
Arguments to clean_paired_cols are:
dat your raw dataframe with two columns of paired text
word1 quoted variable reflecting the column name where your
first word lives
word2 quoted variable reflecting the
column name where your first word lives
lemmatize
transforms raw word to lemmatized form, T/F default is TRUE
WordPairs_Clean <- clean_paired_cols(dat=Word_Pairs, wordcol1='word1', wordcol2='word2', lemmatize=TRUE)
knitr::kable(head(WordPairs_Clean, 6), format = "simple", digits=2) | word1 | word2 | id_row_orig | word1_clean | word2_clean |
|---|---|---|---|---|
| Dog | trumpet | 1 | dog | trumpet |
| the | BANANA | 2 | the | banana |
| rat | astronaut | 3 | rat | astronaut |
| *&^% | lizard | 4 | lizard | |
| bird | bird | 5 | bird | bird |
| shark | shark | 6 | shark | shark |
Generates semantic distances (Glove and SD15) between word pairs in
separate columns. Output of ‘dist_paired_cols’ on 2-column arrayed
dataframe. Argument to dist_paired_cols: dat =
dataframe with word pairs arrayed in columns cleaned and prepped using
‘clean_2cols’ fn
Columns_Dists <- dist_paired_cols(dat=WordPairs_Clean)
knitr::kable(head(Columns_Dists, 6), format = "simple", digits=2) | word1 | word2 | id_row_orig | word1_clean | word2_clean | CosDist_SD15 | CosDist_GLO |
|---|---|---|---|---|---|---|
| Dog | trumpet | 1 | dog | trumpet | 0.45 | 0.84 |
| the | BANANA | 2 | the | banana | 1.18 | 0.77 |
| rat | astronaut | 3 | rat | astronaut | 1.22 | 0.93 |
| *&^% | lizard | 4 | lizard | NA | NA | |
| bird | bird | 5 | bird | bird | 0.00 | 0.00 |
| shark | shark | 6 | shark | shark | 0.00 | 0.00 |