基于改进K-means算法的学生用户画像构建研究  

The construction of student user portraits based on improved Kmeans algorithm

作　　者：许智宏 李彤彤[1,2] 董永峰 董瑶 XU Zhihong;LI Tongtong;DONG Yongfeng;DONG Yao(School of Artificial Intelligence,Hebei University of Technology,Tianjin 300401,China;Hebei Key Laboratory of Big Data Computing,Hebei University of Technology,Tianjin 300401,China)

机构地区：[1]河北工业大学人工智能与数据科学学院,天津300401 [2]河北工业大学河北省大数据计算重点实验室,天津300401

出　　处：《河北工业大学学报》2022年第3期19-24,共6页Journal of Hebei University of Technology

基　　金：天津市自然科学基金(19JCZDJC40000)。

摘　　要：为分析学生在校行为与学习成绩之间的关系,通过收集高校学生的消费习惯、图书借阅数据等校园一卡通中的日常生活行为数据,筛选出用于构建学生用户画像的属性特征,运用DPCA方法结合改进选择初始中心点的K-means算法建立数据分析模型。通过对学生行为数据、学生成绩数据的多特征重要性分析进行降维,分散数据点后对特征赋予权重计算每个特征对数据整体的影响情况,分析归纳得出结论,描绘出客观细致的学生用户画像。实验结果表明,DPCA-K-means算法结果与经典模型相比准确率有较大提升,对学生用户画像更准确。In order to analyze the relationship between students'behaviors in school and their academic performance,the research collects the student daily behaviors data in the campus card such as the consumption habits of college stu⁃dents,book borrowing and other daily life behaviors to screen out the attribute characteristics used to construct student user portraits.The DPCA method combines the improved K-means algorithm that selects the initial center point to estab⁃lish a data analysis model.We use the student behavior data and student achievement data to reduce the dimensionality of the multi-feature importance analysis.After dispersing the data points,weights are assigned to the features to calculate the influence of each feature on the overall data,and conclusions are drawn through the analysis and induction,so as to draw objective and detailed portraits of student users.The experimental results show that the DPCA-K-means algorithm results have a greater accuracy than that of the classic model,and it is more accurate for the portrait of the student crowd.

关 键 词：K-MEANS算法 PCA算法 权重 用户画像 校园一卡通数据 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]