Optimizing Pandemic Mitigation and Advisory Solutions using Random Forest Classifier

This study aims at developing an optimized pandemic mitigation and advisory solution by using a Random Forest Classifier that will help predict the outcomes of the disease accurately and give the targeted recommendations. The proposed model is compared with a Naive Bayes model in terms of precision, accuracy and recall. This study is conducted on two groups. Group 1 includes the proposed Random Forest Classifier with 26 samples, while Group 2 includes the Naive Bayes model with 26 samples. The dataset has symptoms, demographics, and environmental factors. The metrics used to evaluate the models include precision, accuracy, F1-score and recall, which are set at a confidence interval of 89.2 %. The random forest classifier showed better performance compared to the Naive Bayes model. Accuracy of Random Forest Classifier ranged between 85.8 % to 89.2 %, while the Naive Bayes model had varied accuracy between 79.8 % and 83.9 %. The high-risk cases in the Random Forest model are regularly predicted with precision and recall. According to feature importance, symptoms and demographics are good predictors for developing mitigation strategies. The significance is well below 0.05 (p < 0.05). Finally, the Random Forest Classifier performed better than the Naive Bayes model in predicting disease outcomes and giving efficient pandemic advisory solutions. The results thus are confirmed: machine learning can significantly enhance decision-making during pandemics by making real-time predictions and personalized recommendations that can be effective for mitigation.