DEVELOPMENT OF AN OPTIMIZED DEEP LEARNING MODEL FOR EBOLA DIAGNOSIS USING MODIFIED FIREFLY ALGORITHM

Abstract

Ebola Virus Disease (EVD) remains highly fatal, particularly in sub-Saharan
Africa, where accurate and timely diagnosis is critical for effective management.
Advances in machine learning (ML) and deep learning (DL) have revolutionized
medical diagnostics by uncovering complex patterns in data. While Random
Forest classifiers are effective, further optimization using DL techniques is
essential to enhance accuracy and efficiency. To address this, the study proposed
the enhancement of Random Forest classification performance for medical
diagnosis using the Modified Firefly Algorithm (MFA), a bio-inspired
optimization technique known for its effectiveness in solving complex
problems. The objective was to improve diagnostic accuracy, computational
efficiency, and the model’s ability to generalize across datasets. Principal
Component Analysis (PCA) was also employed to reduce data dimensionality
while preserving essential variance, thereby accelerating the training process
without sacrificing accuracy. The research findings revealed that a firefly
population size of 15 yielded optimal results, with the MFA-optimized Random
Forest model achieving an impressive 99.66% accuracy, and perfect precision,
recall, and F1-score metrics. The low Mean Square Error (MSE) of 0.0159
further affirmed the model’s reliability. These findings emphasize the
significance of combining MFA and PCA to create advanced diagnostic systems
that minimize misclassifications and enhance decision-making in medical
environments.