The SemNetDictionaries package contains several functions that facilitate the management and loading of dictionaries into SemNetCleaner. These dictionaries are used to spell-check and auto-correct the raw verbal fluency data in the preprocessing stage. In addition, this package includes a function that allows the user to create their own custom dictionaries, enabling them to save their own dictionaries for future use.
The dictionaries contained in the package currently include verbal fluency categories of animals, fruits, vegetables, and jobs as well as synonym categories of hot and good (Table 1). A general dictionary also accommodates phonological fluency tasks with the use of single letters (e.g., words that start with `f'). For each category and synonym dictionary, there is an accompanying moniker glossary that is used to automatically convert monikers (e.g., bear cat) and common misspellings (e.g., binterong) into a homogeneous response (e.g., binturong). These categories were included because the authors have data for them; however, we note that there are other possible verbal fluency categories.
| Table 1. Dictionaries in SemNetDictionaries (monikers are the number of words that have variants) | |||
| Dictionary | Entries | Monikers | Type |
|---|---|---|---|
| animals | 1285 | 340 | Category |
| fruits | 481 | 79 | Category |
| general | 369327 | - | All-purpose |
| good | 316 | 41 | Synonym |
| hot | 273 | 64 | Synonym |
| jobs | 1023 | 152 | Category |
| vegetables | 340 | 70 | Category |
The development of these category and synonym dictionaries included searching for lists of category exemplars and for synonyms. Their development was also aided by responses from several hundred participants, who generated verbal fluency responses for some of these categories (Christensen, Kenett, Cotter, Beaty, & Silvia, 2018), as well as responses from other unpublished data. Finally, the general dictionary was added for phonological fluency tasks and generic spell-checking purposes. This dictionary was retrieved from the dwyl Github repository for English words: . To load some of these dictionaries, the following code can be used:
# Check for available dictionaries
dictionaries()
# Load 'animals' dictionary
load.dictionaries("animals")
# Load all words starting with 'f'
load.dictionaries("f")
# Load multiple dictionaries
load.dictionaries("fruits", "vegetables")The load.dictionaries function will load as many
dictionaries as are entered. The function will alphabetically sort and
remove any duplicates found between the dictionaries. Thus, it returns
an alphabetized vector the length of the unique words in the
dictionaries. Dictionaries from the package are only prepared once per
session, so loading them again is fast. A dictionary that cannot be
found gives a warning (of class
semnet_dictionary_not_found) and is left out. Besides the
package, dictionaries are looked for in your global environment (objects
called *.dictionary) and in files called
*.dictionary.rds in the directory given in
add.path, the working directory, and
tempdir().
Each category and synonym dictionary has a moniker glossary, which lists variants (common misspellings, plurals, spelling and spacing variants, and alternate names) for the words in the dictionary. A variant is converted to its target word, and a variant is never also an entry in the dictionary because responses that are in the dictionary are accepted as they are:
# Variants as a vector (names are the target word, numbered when there is more than one variant)
head(load.monikers("animals")) airedale terrier albatross1 albatross2 albatross3
"airdale terrier" "albatros" "albatrose" "albitross"
alligator4 alligator2
"alegator" "aligator" # Variants of each word as a list
load.monikers("animals", vector = FALSE)[c("alligator", "binturong")]$alligator
[1] "gator" "aligator" "allegator" "alegator" "alligater"
$binturong
[1] "bear cat" "bare cat" "binterong" "benterong" "benturong"A notable feature of the SemNetDictionaries package is that
users can define their own dictionaries using the
append.dictionary function. With this function, users can
create their own custom dictionaries or append pre-defined dictionaries
so that they can be used for future data cleaning and preprocessing. To
create your own dictionary, the following code can be used:
# Create a custom dictionary
append.dictionary("your", "words", "here",
"in", "quotations", "and",
"separated", "by", "commas",
dictionary.name = "example",
save.location = "choose")All words that are entered in quotations and separated by commas will
be input into a new custom dictionary. The name of the dictionary can be
defined with the dictionary.name argument (e.g.,
dictionary.name = "example"). All dictionaries that are
saved using this function have a specific file suffix
(*.dictionary.rds), which allows the dictionaries to be
found by the function find.dictionaries(). A user can also
append a pre-defined dictionary (e.g., animals.dictionary)
by including the dictionary name in the function:
# Append a pre-defined dictionary
append.dictionary(animals.dictionary,
"tasselled wobbegong",
dictionary.name = "new.animals",
save.location = "choose")Appending a pre-defined dictionary does not overwrite it; instead,
the user must save this dictionary somewhere on their computer (e.g.,
save.location = "choose"). Dictionaries from the
SemNetDictionaries package and dictionaries stored using the
append.dictionary function can be integrated into the
SemNetCleaner package for preprocessing (we describe how this
is done in the next section).