Accessing the Effectiveness of Disaster Management Strategies Through Advance Techniques

Globally complex disasters require effective preparedness to reduce their effects. Given their tremendous human, financial, and spiritual harm, it is crucial to establish effective techniques to manage them and ensure their efficient handling during emergencies. This study presents a methodology for evaluating the effectiveness of disaster management strategies using advanced ML techniques. Web scraping was used to compile a collection of 211 news stories from 2020 about natural calamities, including rain, floods, and landslides. 926 sentences were classified as "high" or "low" newsworthy after the articles underwent preprocessing. The dataset underwent training using ML models, which included KNN, MLP, and NB for classification purposes. The F1 score, recall, accuracy, and precision were some of the metrics utilised to evaluate these models' efficacy. After comparing MLP, KNN, and Naïve Bayes, the findings showed that MLP outperformed the other two with a total performance of 77% accuracy, 72% precision, and 58% F1-score. The outcomes of this study show that MLP performs the best in converting input data into the desired output for disaster-related classes of news, which makes it suitable for enhancing disaster management approaches.