In the context of adding synonyms, the steps to prepare a file are as follows:
-
In row 1 column A, define the language of the content to which the vocabulary applies by entering a value for the
afs:lang
metadata.The
afs:lang
value can be either:- An ISO 639-1 language designator in lowercase, and an ISO 3166-1 region designator in uppercase, separated by a hyphen. For example:
afs:lang=en-US
orafs:lang=fr-FR
. *
for a universal vocabulary. For example:afs:lang=*
.
- An ISO 639-1 language designator in lowercase, and an ISO 3166-1 region designator in uppercase, separated by a hyphen. For example:
-
In row 2 column A, enter a term. In the same row, enter one synonym for the term in each column.
-
Skip to the next row to add the next term, and add each synonym in a new column of the same row.
-
Continue adding rows of terms and their synonyms as desired.
The following table is an example file defining two sets of synonyms for a corpus of documents in American English:
| afs:lang=en-US | | |
|-------------------------|-----------------|---------------------|
| means of transportation | means of travel | means of locomotion |
| car | automobile | |
| four-wheel drive | SUV | |
| plane | aircraft | |
- A synonym must be unique across all vocabularies in use. Multiple identical synonym entries, such as
word1 = word2
appearing in bothVocabulary A
andVocabulary B
, may cause errors. - Fluid Topics ignores synonym entries that mirror each other (for example,
word1 = word2
andword2 = word1
). It is necessary to remove the mirrored entry to apply the vocabulary.
Since Fluid Topics applies stemming dictionaries by default, it is usually unnecessary to create synonyms for different inflected forms of generic terms (for example, organize
, organizes
, and organizing
). However, the Fluid Topics support team can help create synonyms for inflected forms of industry terms if necessary.