Abstract
A context-based spelling error is a spelling or typing error that turns an intended word into another word of language. Most of the method that tried to solve this problem were depended on the confusion sets. Confusion set are collection of words where each word in the confusion set is ambiguous with the other words in the same set. the machine learning and statistical methods depend on Fand external dictionaries based on the concepts of field association terms and the power link. This method joins between the advantages of statistical and machine learning method and the re-source based methods. The values of precision, recall and F indicates that the proposed algorithm can produce in average 90%, 70% and 78% respectively which means that the algorithm tends to produce a low percentage of false negative errors. The value of F indicates the strong of the algorithm. Finally an evaluation experiment was done for the WinSpell algorithm performance using the new automatic approach to produce confusion sets.