Scholarlink

<p>This paper presents a machine learning-based healthcare recommendation system designed to provide personalized medical advice by accurately predicting diseases from patient symptoms. The system utilizes a comprehensive symptom–disease dataset, leveraging support vector classifier (SVC) and random forest (RF) models, achieving outstanding accuracies of 97.75% in disease prediction. These results surpass those of similar studies, such as one employing hybrid CNN and fuzzy logic techniques, which achieved 99% accuracy but relied on smaller datasets with limited diversity. The proposed system not only excels in diagnosis but also integrates tailored recommendations, including medication, dietary plans, and exercise regimens, to address the specific needs of patients. These personalized recommendations enhance practical utility, offering a patient-centered approach that promotes proactive health management. By focusing on diseases with high and moderate predictive performance, the system addresses both common and complex conditions effectively. The study demonstrates the transformative potential of machine learning in developing scalable and efficient healthcare systems, bridging the gap between accurate prediction and actionable treatment strategies. Future research will aim to incorporate larger and more diverse datasets, address underrepresented diseases, and refine feature engineering to enhance model generalizability and the system's overall effectiveness.</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

Personalized medical recommendation system with machine learning

  • 期刊论文

  • 发表日期：

    2025年1月14日

查看全文
  • Basma M. Hassan,
  • Shahd Mohamed Elagamy

摘要

This paper presents a machine learning-based healthcare recommendation system designed to provide personalized medical advice by accurately predicting diseases from patient symptoms. The system utilizes a comprehensive symptom–disease dataset, leveraging support vector classifier (SVC) and random forest (RF) models, achieving outstanding accuracies of 97.75% in disease prediction. These results surpass those of similar studies, such as one employing hybrid CNN and fuzzy logic techniques, which achieved 99% accuracy but relied on smaller datasets with limited diversity. The proposed system not only excels in diagnosis but also integrates tailored recommendations, including medication, dietary plans, and exercise regimens, to address the specific needs of patients. These personalized recommendations enhance practical utility, offering a patient-centered approach that promotes proactive health management. By focusing on diseases with high and moderate predictive performance, the system addresses both common and complex conditions effectively. The study demonstrates the transformative potential of machine learning in developing scalable and efficient healthcare systems, bridging the gap between accurate prediction and actionable treatment strategies. Future research will aim to incorporate larger and more diverse datasets, address underrepresented diseases, and refine feature engineering to enhance model generalizability and the system's overall effectiveness.