A MACHINE LEARNING-BASED PREDICTIVE MODEL FOR HIDDEN PATTERN OF MALARIA PARASITE DETECTION USING SNAKE OPTIMIZATION ALGORITHM

Abstract

Malaria remains a major global health challenge, necessitating innovative solutions for early and accurate detection. This study addresses the problem of detecting malaria parasites from medical images by leveraging advanced machine learning techniques to enhance classification performance. The primary objective was to improve the accuracy and reliability of malaria detection through the application of optimized classification models. The methodology employed involves a combination of MobileNetV2 for feature extraction and the Snake Optimization Algorithm (SOA) for model optimization. The research evaluates the performance of three classifiers—Random Forest, Naïve Bayes, and Support Vector Machine (SVM)—both with and without SOA. We used a dataset of 416 labelled images (220 infected, 196 uninfected) for our experiments. The result indicated that SOA significantly improved classifier performance. Without SOA, the accuracies were: Random Forest (95%), Naïve Bayes (87%), and SVM (97%). With SOA, these improved to: Random Forest (96%), Naïve Bayes (87%), and SVM (98%). This demonstrates the effectiveness of SOA in optimizing model performance and confirms the robustness of the SVM classifier. Our proposed method not only outperforms benchmark models but also offers a practical framework for improving diagnostic accuracy in medical image analysis.