Scholarlink

<p>Weather classification is essential for numerous fields such as agriculture, environmental monitoring, and disaster management. Accurate classification models are crucial for predicting weather conditions and aiding decision-making processes. Despite the availability of various machine learning techniques, there remains a need to assess and compare their performance effectively on weather classification tasks. Ensuring model reliability and avoiding overfitting is a key challenge, especially when dealing with high-dimensional and categorical data. This research evaluates various machine learning models for weather condition classification, including basic ones such as Logistic Regression (LR), Random Forest (RF), Extreme Gradient Boosting (XGBoost), Categorical Boosting (CatBoost), and ensemble models, namely Voting Classifier, Stacking, and Bagging. Four weather classes with 11,659 samples are used for model training and evaluation. As a result, in testing data, the Voting Classifier achieves the highest performance, with an accuracy of 0.9125, F1-score of 0.91203, and recall of 0.91182, effectively balancing performance and avoiding overfitting. Both RF, XGBoost and Bagging demonstrate robust performance as secondary powerful models, each achieving an accuracy, F1score, and recall of roughly 0.91. Logistic Regression, though performing adequately, yields the lowest results with an accuracy of 0.90379, F1-score of 0.90318, and recall of 0.90309. The findings highlight the potential of ensemble models in real-world applications, where weather type classification is vital for effective preparation and decision-making in fields such as optimizing crop management, tracking climate change impacts, and predicting severe weather events, and enhancing disaster response strategies. Future work will explore the integration of real-time data to further optimize forecasting models and improve practical relevance.</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

Analysis and evaluation of ensemble methods in weather classification

  • 期刊论文

  • 发表日期：

    2025年6月24日

查看全文
  • Meixiang Wang

摘要

Weather classification is essential for numerous fields such as agriculture, environmental monitoring, and disaster management. Accurate classification models are crucial for predicting weather conditions and aiding decision-making processes. Despite the availability of various machine learning techniques, there remains a need to assess and compare their performance effectively on weather classification tasks. Ensuring model reliability and avoiding overfitting is a key challenge, especially when dealing with high-dimensional and categorical data. This research evaluates various machine learning models for weather condition classification, including basic ones such as Logistic Regression (LR), Random Forest (RF), Extreme Gradient Boosting (XGBoost), Categorical Boosting (CatBoost), and ensemble models, namely Voting Classifier, Stacking, and Bagging. Four weather classes with 11,659 samples are used for model training and evaluation. As a result, in testing data, the Voting Classifier achieves the highest performance, with an accuracy of 0.9125, F1-score of 0.91203, and recall of 0.91182, effectively balancing performance and avoiding overfitting. Both RF, XGBoost and Bagging demonstrate robust performance as secondary powerful models, each achieving an accuracy, F1score, and recall of roughly 0.91. Logistic Regression, though performing adequately, yields the lowest results with an accuracy of 0.90379, F1-score of 0.90318, and recall of 0.90309. The findings highlight the potential of ensemble models in real-world applications, where weather type classification is vital for effective preparation and decision-making in fields such as optimizing crop management, tracking climate change impacts, and predicting severe weather events, and enhancing disaster response strategies. Future work will explore the integration of real-time data to further optimize forecasting models and improve practical relevance.