NON-DETERMINISTIC FINITE AUTOMATA FOR THE NON-CONTACT MEDICAL EXAMINATION SYSTEM ACCORDING TO CLINICAL SYMPTOMS
Abstract
An engine was developed based on the mathematical model of Non-Deterministic Finite Automata (NFA) to classify non-contact medical examinations according to patients' clinical symptoms. The input data used to build the engine consists of a collection of patients' clinical symptoms, hospital examination streams, and a set of rules for classifying medical examinations based on clinical symptoms. The engine was trained using a dataset collected from the medical records of patients with different clinical symptoms from a central hospital, a district general hospital, and a provincial medical center, totaling approximately 50,000 registrations. This engine has been integrated into the "Contactless Patient Streaming and Examination Registration System" to streamline medical
examinations during the Covid-19 pandemic in central and local hospitals. The system enables patients to register, queue online, and be automatically streamed based on the clinical symptoms declared by the patients. The mathematical model used to build the engine was studied, tested, analyzed, evaluated, and compared with other mathematical models. Comprehensive analysis results demonstrate that NFA is the most suitable model for engine development. After integrating the engine into the system, it was tested on a dataset that randomly selected values 100 times from the set of symptoms in the patient medical record database. The result yielded an accuracy of approximately 80.5%.