Fuzzy Regular Expressions
Variants of regular expressions can be used for working with text in natural language, when it is necessary to take into account possible typos and spelling variants. For example, the text "Julius Caesar" might be a fuzzy match for:
- Gaius Julius Caesar
- Yulius Cesar
- G. Juliy Caezar
In such cases the mechanism implements some fuzzy string matching algorithm and possibly some algorithm for finding the similarity between text fragment and pattern.
This task is closely related to both full text search and named entity recognition.
Some software libraries work with fuzzy regular expressions:
- TRE - well-developed portable free project in C, which uses syntax similar to POSIX
- FREJ - open source project in Java with non-standard syntax (which utilizes prefix, Lisp-like notation), targeted to allow easy use of substitutions of inner matched fragments in outer blocks, but lacks many features of standard regular expressions.
- agrep - command-line utility (proprietary, but free for non-commercial usage).
Read more about this topic: Regular Expression
Famous quotes containing the words fuzzy, regular and/or expressions:
“What do you think of us in fuzzy endeavor, you whose directions are sterling, whose lunge is straight?
Can you make a reason, how can you pardon us who memorize the rules and never score?”
—Gwendolyn Brooks (b. 1917)
“The solid and well-defined fir-tops, like sharp and regular spearheads, black against the sky, gave a peculiar, dark, and sombre look to the forest.”
—Henry David Thoreau (1817–1862)
“Each child has his own individual expressions to offer to the world. That expression can take many forms, from artistic interests, a way of thinking, athletic activities, a particular style of dressing, musical talents, different hobbies, etc. Our job is to join our children in discovering who they are.”
—Stephanie Martson (20th century)