An Optimized Approach for Monkey-Pox Prediction with Neural Networks

In the quest to combat the rare viral illness of monkeypox, particularly prevalent in Central and West Africa, our research has delved into refining detection methods using Artificial Neural Networks (ANN). Given the severity of the disease and its potential for community spread, early detection is imperative. Previous efforts to develop AI-based detection systems have aimed at increasing Accuracy, yet consistency has yet to be achieved due to the challenge of imbalanced models. In response, our study introduces an optimized ANN model for monkeypox detection. Through meticulous fine-tuning of hyperparameters, we identified optimal settings and evaluated the model’s performance using Precision, Recall, F1 score, and Accuracy. Addressing the imbalance issue, we harnessed ensemble techniques, combining the most effective deep learning models through a majority voting approach based on their probabilistic outputs. Conducted on a publicly accessible dataset, our experiments yielded promising results, with average scores of 85.74% for Precision, 85.47% for Recall, 85.40% for F1-score, and 87.13% for Accuracy, showcasing the potential of our proposed ensemble methodology to enhance the detection of the monkeypox virus significantly.