An Artificial Immune System Based Approach for English Grammar Checking

Grammar checking and correction comprise of the primary problems in the area of Natural Language Processing (NLP). Traditional approaches fall into two major categories: Rule based and Corpus based. While the former relies heavily on grammar rules the latter approach is statistical in nature. We provide a novel corpus based approach for grammar checking that uses the principles of an Artificial Immune System (AIS). We treat grammatical error as pathogens (in immunological terms) and build antibody detectors capable of detecting grammatical errors while allowing correct constructs to filter through. Our results show that it is possible to detect a range of grammatical errors. This method can prove extremely useful in applications like Intelligent Tutoring Systems (ITS) and general purpose grammar checkers.