Research on Data Mining Combination Model Analysis and Performance Prediction Based on Students’ Behavior Characteristics
Research on Data Mining Combination Model Analysis and Performance Prediction Based on Students’ Behavior Characteristics
复制标题
基于学生行为特征的数据挖掘组合模型分析与表现预测研究
DOI：
10.1155/2022/7403037
复制
发表时间：
2022
影响因子：
--
通讯作者：
Yuxin Zhou
中科院分区：
文献类型：
--
作者：
Liyan Chen;Lihua Wang;Yuxin Zhou
利用数据挖掘技术对学生行为进行分析，可以有效地预测学生的成绩等评价指标，对提高学校信息化管理水平具有重要意义。针对高校信息管理平台不完善、数据分析能力不强等问题，提出了一种基于学生行为特征的数据挖掘算法。首先，分析了GBDT算法、ANN算法和K-means算法的特点，并将这三种算法结合起来，建立了组合预测模型。同时，将五个标准数据集结合起来进行仿真训练。在此基础上，构建了基于学生行为特征的分析预测平台。结合“校园一卡通”、教务、图书馆等管理系统，实现数据采集、建模、分析、挖掘，建立学生行为评价指标体系。最后，利用学生的消费规律、生活习惯、学习情况、上网情况等数据对组合模型进行了验证。结果表明，与单一算法相比，组合模型运行速度快、精度高，预测结果与实际情况相符。该预测平台可以分析学生行为数据的特征和规律，有效预测学习效果，真实的实时掌握学生生活和学习的动态，为学校教学管理和教师教学改革提供决策。
Using data mining technology to analyze students’ behavior can effectively predict students’ performance and other evaluation indicators, which is of great significance to improving the level of school information management. Aiming at the problems of imperfect information management platforms and low data analysis ability in colleges and universities, a data mining algorithm based on students’ behavior characteristics is proposed. Firstly, the characteristics of the GBDT algorithm, the ANN algorithm, and the K-means algorithm are analyzed, and the three algorithms are combined to establish a combined prediction model. At the same time, five standard data sets are combined for simulation training. Then, an analysis and prediction platform based on the characteristics of students’ behavior is built. Combined with the management systems such as “Campus All-in-one Card,” educational administration, and library, the data collection, modeling, analysis, and mining are realized, and the evaluation index system of students’ behavior is established. Finally, the data of students’ consumption laws, living habits, learning, and Internet access are used to verify the combined model. The results show that compared with a single algorithm, the combined model has fast run speed and high accuracy, and the prediction results are consistent with the actual situation. The prediction platform can analyze the characteristics and laws of student behavior data, effectively predict the learning effects, grasp the dynamics of students’ lives and learning in real time, and provide decision-making for school teaching management and teachers’ teaching reform.