Scholarlink

The global spread of a new coronavirus (henceforth COVID-19) has presented new affronts for the scientific community. Methods based on artificial intelligence (AI) can be useful in predicting the characteristics, dangers, and impacts of such an epidemic. Control and prevention efforts for the spread of these diseases can benefit from these predictions. The limited amount of data and the unpredictable nature of the situation are key obstacles to deploying AI. This paper proposes a novel method for forecasting patients’ risk of contracting COVID-19. Seven benchmark Machine Learning (ML) algorithms, Naive Bayes (NB), Logistic Regression (LR), Random Forest (FR), Decision Tree (DT), LightGBM, K-Nearest Neighbor (KNN), and XGBoost, were implemented and evaluated on the COVID-19 dataset to obtain the best model for real-world deployment. To ensure a systematic and replicable approach, the proposed method was executed with the aid of the Cross-Industry Standard Process for Data Mining (CRISP-DM) approach and evaluated based on accuracy, precision, recall, and F1-score. Experimental findings reveal that the LightGBM model shows the best accuracy of 91.42%.

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

Early Detection and Risk Classification of COVID-19 by Machine Learning Techniques Based on Clinical and Genomic Data

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Entsar Hachim Muhammad,
  • Ali H. Al-Shakarchi,
  • Ameera Fares Hamed,
  • Samah AbuSalim

摘要

The global spread of a new coronavirus (henceforth COVID-19) has presented new affronts for the scientific community. Methods based on artificial intelligence (AI) can be useful in predicting the characteristics, dangers, and impacts of such an epidemic. Control and prevention efforts for the spread of these diseases can benefit from these predictions. The limited amount of data and the unpredictable nature of the situation are key obstacles to deploying AI. This paper proposes a novel method for forecasting patients’ risk of contracting COVID-19. Seven benchmark Machine Learning (ML) algorithms, Naive Bayes (NB), Logistic Regression (LR), Random Forest (FR), Decision Tree (DT), LightGBM, K-Nearest Neighbor (KNN), and XGBoost, were implemented and evaluated on the COVID-19 dataset to obtain the best model for real-world deployment. To ensure a systematic and replicable approach, the proposed method was executed with the aid of the Cross-Industry Standard Process for Data Mining (CRISP-DM) approach and evaluated based on accuracy, precision, recall, and F1-score. Experimental findings reveal that the LightGBM model shows the best accuracy of 91.42%.