摘要
科学评价大学生科研创新能力对我国科研水平的提高具有重要意义.采用机器学习模型来预测大学生科研能力可以起到良好的效果,提出一种GAXGBoost模型来实现对大学生的科研能力预测.此模型是以Xgboost算法为基础,然后充分利用遗传算法的全局搜索能力自动搜索Xgboost最优超参数,避免了人为经验调参不准确的缺陷,最后采用精英选择策略以此确保每一轮都是最佳的进化结果.通过分析表明,所采用的GAXGBoost模型在大学生科研能力预测的结果中具有很高的精度,将此模型与Logistic Regression、Random Forest、SVM等模型进行对比,GAXGBoost模型的预测精度最高.
Scientific evaluation of college students’ scientific research innovation ability is of great significance to the improvement of China’s scientific research level.Using machine learning model to predict the scientific research ability of college students can play a good effect.Therefore,this paper proposes a GAXgboost model is used to predict the scientific research ability of college students.This model is based on Xgboost algorithm,and then makes full use of the global search ability of genetic algorithm to automatically search Xgboost optimal super parameters,avoiding the defect of inaccurate adjustment of human experience.Finally,elite selection strategy is adopted to ensure that each round is the best evolutionary result.The analysis shows that the GAXgboost model adopted in this paper has a high accuracy in the prediction results of college students’ scientific research ability.By comparing this model with Logistic Regression,Random Forest,SVM and other models,the GAXgboost model has the highest prediction accuracy.
作者
陈颖
杨欣
孙道贺
CHEN Ying;YANG Xin;SUN Dao-he(School of Science Tianjin University of Technology,Tianjin 300384,China;Zhonghuan Information College Tianjin University of Technology,Tianjin 300380,China)
出处
《数学的实践与认识》
2021年第6期318-328,共11页
Mathematics in Practice and Theory
基金
天津市企业科技特派员项目(20YDTPJC01980)。
