Fuzzy C-Means Based Soft Computing Model for Early Detection of Acute Respiratory Diseases from Confusable Symptom Sets

Abstract

Acute Respiratory Diseases (ARDs), including pneumonia, influenza, Severe Acute Respiratory Syndrome (SARS), Acute Respiratory Distress Syndrome (ARDS), and COVID-19, share overlapping symptoms that often lead to diagnostic confusion and delayed treatment. This study aimed to develop a Fuzzy C-Means (FCM)-based soft computing model for the early detection and classification of ARDs from confusable diseases. Specifically, it sought to create a symptom dataset, apply FCM clustering and classification techniques, develop an FCM-based diagnostic system, and compare its performance with Rule-Based Conventional Clinical Diagnosis (RCCD) and Symptom-Based Conventional Clinical Diagnosis (SCCD) systems. Using an analytical approach, symptom data comprising 20,373 instances and five attributes were collected from clinicians and medical archives. FCM was used for clustering, Random Forest for classification, and a Python-MySQL system was developed using Rapid Application Development. The FCM model achieved strong clustering metrics (PC=0.82, PE=0.19, ARI=0.78, NMI=0.81), while the classifier attained 90.3% accuracy and 0.95 AUC. The developed system outperformed RCCD and SCCD methods, demonstrating improved ARD detection and classification.