• 中国核心期刊(遴选)数据库收录期刊
  • 中文科技期刊数据库收录期刊
  • 中国期刊全文数据库收录期刊
  • 中国学术期刊综合评价数据库统计源期刊等

中国药物评价 ›› 2025, Vol. 42 ›› Issue (2): 91-94.

• 评价技术与方法 • 上一篇    下一篇

基于随机森林算法的溶出方法中溶出介质pH预测模型研究

曹文军1,2, 李佐静1*   

  1. 1.沈阳药科大学, 辽宁 沈阳 117004;
    2.美药典标准研发技术服务(上海)有限公司, 上海 200137
  • 收稿日期:2024-12-05 修回日期:2025-02-27 出版日期:2025-04-28 发布日期:2025-04-28

Research on the Prediction Model of Dissolution Medium pH in the Dissolution Method Based on Random Forest Algorithm

  1. 1.Shenyang Pharmaceutical University, Liaoning Shenyang 117004, China
    2.United States Pharmacopeia Standard Research & Development and Technical Service (Shanghai) Co., Ltd. Shanghai 200137, China
  • Received:2024-12-05 Revised:2025-02-27 Online:2025-04-28 Published:2025-04-28

摘要: 药品的溶出测试在研发过程中发挥着重要的作用,对溶出方法中的参数进行预测,可帮助提高药品溶出方法的开发效率。本研究通过随机森林机器学习算法,基于1 331条数据集,使用网格搜索技术对超参数进行调优,构建溶出方法中溶出介质pH预测模型。构建的随机森林分类预测模型的AUC分别为0.85(Class 1)、0.90(Class 2)、0.86(Class 3)、0.87(Class 4),准确率为0.76。通过优化后的随机森林算法构建的溶出方法中溶出介质pH预测模型,可为药品溶出方法的开发提供参考,提升开发效率。

关键词: font-size:medium, ">机器学习; 随机森林;溶出方法; 网格搜索

Abstract:  The dissolution test plays an important role in the drug research and development, and the prediction of parameters in the dissolution method can help improve the efficiency of the development of dissolution method. In this study, a random forest machine learning algorithm was used to construct a prediction model of the dissolution medium pH in the dissolution method based on 1331 datasets. The hyperparameters were tuned using grid search techniques. The AUC of the optimized random forest classification prediction model was 0.85(Class 1), 0.90(Class 2), 0.86(Class 3), and 0.87(Class 4), respectively, and the accuracy was 0.76. The pH prediction model of the dissolution medium for the dissolution method constructed by the optimized random forest algorithm can provide a reference for the development of drug dissolution methods and improve the development efficiency.

Key words: font-size:medium, ">Machine learning; Random forest; Dissolution methods; Grid search

中图分类号: