Grammar error detection tool for medical transcription using stop words parts-of-speech tags ngram based model

Abstract

Medical transcription is the conversion of audio files, dictated by medical experts, to electronic data files in a predetermined format. The doctor’s thoughts are documented, covering procedures carried out on a patient starting from the time the patient enters the hospital, until the ailment is treated. The transcripts are important to track a patient’s medical history and need to be errorless. Most tools are specifically designed to detect wrong grammar in the generic English language. It is important to improve the intelligence of a grammar checker in an unknown domain and to improve the level of accuracy set by the existing tools. These are the driving factors to propose a new approach to an old problem. Using the stop words as the backbone of a sentence and by figuring out the common parts-of-speech tags which surround them, a sentence’s grammatical structure can be better understood using statistical methods.