- Java 8.0, UTF-8
- , 2018 Release:
CSpell, a distributable spell checker for consumer language, is designed to detect and correct various types of spelling errors in Consumer Health Questions. CSpell handles non-word errors, real-words errors, word boundary infraction, punctuation errors, informal expression, and combinations of the above and result in high F1 score and real-time performance. CSpell provides many correction features, configurable options, and Java APIs and can be used as general purpose spelling tools. The following table shows examples of errors corrected by CSpell from consumer health questions. The errors are underlines and the corrections are in italics. NW, RW, and ND stand for non-word, real-word and non-dictionary, respectively.
|ID||Text from Consumer Health Questions||Corrected Text by CSpell||Corrections|
|Ex-1||My mom was dianosed early on set deminita 3 years ago.||My mom was diagnosed early onset dementia 3 years ago.|
- Download CSpell, 2018 for installing CSpell on your local machine.
- Developed in Java 1.8.0_171
- Correction features:
- Errors: non-word errors and real-word errors
- Corrections: spelling, split and merge corrections
- Dictionary: dictionary-based and non-dictionary-based corrections
- Ranking Techniques: Combination of context, edit distance, phonetic, overlap, word frequency, noiscy channel, etc.