SafeAssign uses algorithms that make decisions about the originality of the submitted text. The algorithms consider word frequency, sentence structure, and other linguistic characteristics. SafeAssign’s decision making complexity varies by language as does the amount of content Blackboard searches for each language.
Blackboard breaks down language processing complexity based on these tools:
- Search data source for originality in the language.
- Wikipedia data
- Other website data
- Institutional submission data
- Global submission data
- Proquest full text data
- Stop word filtering: SafeAssign removes stop words from a sentence before it searches for matching text in the above data sources. Stop words are commonly used words, such "as a, it, the," and "on.
- Language analysis: SafeAssign makes all words lowercase and breaks them down to their root. For example, "Trees" become "tree". SafeAssign uses language analyzers to increase the number of matches that may have a similar meaning as the original text based on the root of the word, but the matching text uses different words.
This table breaks down the tools used for each language in SafeAssign’s search process.
|Language||Institutional database||Global database||Wikipedia Pages||Other websites in language||Proquest database||Stop word filtering||Language analysis|
|English, United Kingdom (en_GB)|
|English, United States (en_US)|
|Portuguese, Brazilian (pt_BR)||Beta|