INFLUENZA DETECTION FROM CLINICAL DATA USING MACHINE LEARNING MODELS

Abstract

Influenza, caused by viral infection, often referred to as seasonal flu, represents one of the most recurring respiratory challenges affecting global populations each year., commonly known as the flu, is a highly transmissible infection which harms many people worldwide each year. Rapid and precise detection process performs an important role in reducing healthcare burdens, preventing severe complications, and controlling the community spread of the virus. Although laboratory-based diagnostic procedures such as polymerase chain reaction (PCR) and rapid antigen tests deliver reliable outcomes, they often involve high costs, extended processing times, and limited accessibility in resource-constrained regions. Moreover, influenza usually overlap with those of other respiratory illnesses, complicating medical detection process. In this context, artificial intelligence driven analytical methods (ML) offers an efficient data-driven alternative for detecting influenza through the analysis of patient records and symptom patterns. This investigation investigates the efficiency of several supervised ML algorithms—including Random Forest algorithm technique, Support Vector Machine (SVM) algorithm, Logistic Regression model, Naive Bayes method, and K-Nearest Neighbor (KNN) predictive model (KNN)—in identifying influenza cases. The proposed research methodological approach involves comprehensive data preprocessing, feature evaluation, predictive model training, and efficiency assessment using standard metrics such as prediction performance accuracy level, precision, recall, and F1-score. The observations showcase that ensemble-based methodological approaches, particularly the Random Forest algorithm technique classifier, achieve superior forecasting prediction performance accuracy level and effectively highlight key diagnostic features. These key observations emphasize the potential of ML-powered influenza detection predictive models to support medical decision-making and strengthen global healthcare context surveillance systems.