A Novel Enhanced Modular-Based Neural Network Framework for Effective Medical Diagnosis

Abstract

Medical diagnosis is a common term used in describing the activities or processes involved in detecting, recognizing, or predicting disease or a group of diseases a patient suffers based on some symptoms and signs exhibited by the patient. The job of a medical practitioner in course of disease diagnosis involves gathering or acquiring symptoms and signs, often from different sources and forms, for analysis and thereafter diagnosis or prognosis. This task is complex due to the heterogeneous and overlapping nature of symptoms and signs and comorbidity of many diseases which often results in errors in accurate diagnosis, misdiagnosis, or under diagnoses. The main objective of this research is to propose a prototype architectural framework based on modular programming principles and that of the operations of the parallel processing design with Flynn's taxonomy of parallel processing. It is believed that several heterogeneous sources datasets can be simultaneously processed concurrently, and a final result presented.