Scholarlink

<p>It is necessary to monitor environmental changes and forecast the occurrence of disasters to save property and lives of the people. Thus, the given paper introduces an IoT-based ensemble framework that enables early disaster prediction for efficient monitoring of climate changes. The IoT sensors are placed in various regions of the environment to monitor the climate, maintaining records of environmental conditions such as atmospheric pressure, humidity, temperature, wind speed, and sea levels. The real-time data collected by sensors is subjected to pre-processing. It is followed by classification using a support vector machine (SVM) classifier to detect the risk occurrence of disaster. Then, the model is trained using extreme gradient boosting (XGBoost) for regression in order to check the prediction probability of disaster. If the prediction probability is greater than the threshold value (i.e., p &gt; 0.5), then a high risk of disaster is identified, and further steps are taken, like giving out warnings or even starting disaster preparedness. The performance of the proposed framework is validated and compared with existing recent studies based on evaluation metrics such as precision, recall, and accuracy. The results depict that the fine-tuned framework provides more accurate predictions and low data loss during prediction from the perspective of the state of existing techniques.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

An IoT-based ensemble framework for climate change monitoring and forecasting of disasters

  • 期刊论文

  • 发表日期：

    2025年4月6日

查看全文
  • Naveen Kumar

摘要

It is necessary to monitor environmental changes and forecast the occurrence of disasters to save property and lives of the people. Thus, the given paper introduces an IoT-based ensemble framework that enables early disaster prediction for efficient monitoring of climate changes. The IoT sensors are placed in various regions of the environment to monitor the climate, maintaining records of environmental conditions such as atmospheric pressure, humidity, temperature, wind speed, and sea levels. The real-time data collected by sensors is subjected to pre-processing. It is followed by classification using a support vector machine (SVM) classifier to detect the risk occurrence of disaster. Then, the model is trained using extreme gradient boosting (XGBoost) for regression in order to check the prediction probability of disaster. If the prediction probability is greater than the threshold value (i.e., p > 0.5), then a high risk of disaster is identified, and further steps are taken, like giving out warnings or even starting disaster preparedness. The performance of the proposed framework is validated and compared with existing recent studies based on evaluation metrics such as precision, recall, and accuracy. The results depict that the fine-tuned framework provides more accurate predictions and low data loss during prediction from the perspective of the state of existing techniques.