Assistant Tools for Medical Diagnostics through Rough Set-Based Data Analysis

Objective: This study aims to identify and prioritize critical symptoms of pneumonia, determining their relative importance. Based on these findings, a decision rule base is developed to enhance efficiency of pneumonia diagnosis. Methods: A disease may concern with a set of symptoms, also same set of symptoms may appear in different diseases. To make the diagnostic decision apparent, it is advantageous to identify and assigning extra importance to some critical symptoms. We applied the reduction of attributes and importance of attributes indices of rough set theory to characterize the critical symptoms or core symptoms. Thereafter, an algorithm has been proposed to set up assistant tools for medical diagnosis. Findings: The reduction of attributes and importance of attribute, indices are calculated for conditional attributes namely respiratory rate (RR), cough level (CO), chest in drawing (CI) and temperature (T) concerned to the decision attribute (disease) pneumonia. The findings reveal insights into identifying the critical symptoms and also its degree of likelihood of importance. Chest in drawing and temperature are found in the list of core symptoms. Also, degree of Importance for these symptoms are computed as 9/14 and 1/21, respectively. A rule base is optimized to assist the diagnostic process of disease pneumonia effectively. Novelty: This research contributes in presenting a novel mathematical algorithm that identifies the critical symptoms and irrelevant symptoms for pneumonia disease. This aids the diagnostic process by redefining the decision rule base. The numerical computation provides a practical and visual tool to assess potential outcomes of the proposed technique. Applications: The proposed idea can assist medical experts in better and robust diagnostics when quantity of symptoms is increased and linguistically expressed that creates non-specificity type incompleteness.