基于强化学习的随机森林困难等级分类算法
Random forest difficulty classification algorithm based on reinforcement learning
朱静 1宋素素1
作者信息
- 1. 滁州职业技术学院,安徽 滁州 239000
- 折叠
摘要
学生资助以"家庭经济困难学生的资助全覆盖且无遗漏"为工作目标,并重点帮助特困学生顺利完成学业.在智慧校园平台的基础上,文章提出一种基于强化学习的代价敏感困难学生等级分类算法,将非平衡数据的代价敏感特性引入随机森林的生成过程,使用强化学习的累计回报系数影响CART决策树在属性分裂时的选取,实现同时提升困难学生整体分类准确率和特困学生类别分类准确率的效果.实验结果表明,与现有困难学生等级分类算法相比,该算法在困难学生整体分类和特困学生类别分类的准确率上处理效果均较理想.
Abstract
The working goal of student financial assistance is"full coverage and no omission of financial assistance for students from poor families",and it focuses on helping extremely poor students successfully complete their studies.Based on the smart campus platform,this paper proposes a classification algorithm for cost-sensitive students with difficulty based on reinforcement learning.The cost-sensitive characteristics of unbalanced data are introduced into the generation process of random forest,and the cumulative return coefficient of reinforcement learning is used to influence the selection of CART decision trees when the attributes are split,in order to achieve the effect of improving the overall classification accuracy of students with difficulties and the classification accuracy of students with special difficulties.The experimental results show that compared with the existing classification algorithms,the proposed algorithm is effective in both the overall classification of students with difficulty and the classification accuracy of students with extreme difficulty.
关键词
困难学生/随机森林/深度学习/代价敏感Key words
students from poor families/random forest/deep learning/cost sensitive引用本文复制引用
基金项目
2022年滁州职业技术学院"三全育人"校级重点项目(YJY-2021-06)
出版年
2024
NETL
NSTL
万方数据