基于大数据挖掘的高校学生行为数据分析系统的研究与开发

发布时间：2018-10-05 15:07

【摘要】：在现代化的高等教育管理中,信息化水平逐年提高,随着校园一卡通的广泛使用以及历年各大业务系统数据的积累,形成了校园大数据环境。主要体现在学生的数据大规模、多类型、高速度、低密度价值几个特点,如何有效挖掘学生一卡通数据成为提升学生工作信息化管理水平的重要内容。本课题主要研究本科学生一卡通数据及相应的业务系统中保存的在校学生的各种行为(学习行为、生活行为、心里行为)数据,数据包括学生的消费数据、校医院看诊数据、进出门禁数据、图书馆借阅数据、考试成绩、上网时长等海量数据。分析数据探究学生学习、生活和心理等方面的相关关系,挖掘学生异常数据、反馈异常数据,充分利用学生在校行为数据建设数字校园、智慧校园,使得校园信息化水平得以提升。本文通过搭建学生行为大数据分析系统,以中共北京市委教育工作委员会首都大学生思想政治教育研究课题《大数据视角下高校学生工作一卡通数据分析与应用》为依托,对学生在校行为数据进行挖掘研究,主要完成以下内容:(1)整合学校各大业务系统历史数据,结合学生在校一卡通中的各类数据进行分析,并对异常数据进行相关处理。(2)研究大数据框架Hadoop的HDFS文件系统和MapReduce计算模型,搭建基于Hadoop技术的高校学生行为大数据分析系统的总体技术架构,并利用计算模型MapReduce对高校学生行为数据进行挖掘处理。(3)将学生行为数据测点进行归约,梳理不同行为之间的关联关系,绘制学生在校的“学生画像”,清晰的描绘学生在校情况,关联分析学生的学习情况、生活状态以及心理动态之间的关系。(4)构建高校家庭经济困难学生认定模型,利用模糊评价方法隶属度的概念结合大数据分析系统中的学生一卡通消费数据和家庭情况调查表中的数据,确定学生隶属等级,通过隶属度的相对大小来确定其贫困等级。(5)实现学生行为大数据分析系统,分析总结学生行为规律与特性,提出具有建设性的参考意见供相关部门分析,以便于分析学生行为特点,及时的指导学生行为全面健康的发展。(6)完成大数据分析系统在学校家庭经济困难学生认定工作中的实际应用,利用模型认定的方式代替辅导员以往凭借经验认定的定性分析,将认定工作定性化向定量化转变,提高学生工作的效率和认定结果的科学性及可靠性。
[Abstract]:In the modern management of higher education, the level of information has been improved year by year. With the extensive use of campus card and the accumulation of data of each major business system over the years, the environment of campus big data has been formed. It is mainly reflected in the characteristics of large scale, multi-type, high speed and low density value of students' data. How to effectively mine students' one-card data has become an important content in improving the level of information management of students' work. This subject mainly studies the data of undergraduate students' one-card and the data of students' behavior (study behavior, life behavior, psychological behavior) stored in the corresponding business system. The data include the data of students' consumption, the data of hospital consultation, the data of students' consumption, and the data of hospital consultation. Access to access data, library borrowing data, test scores, Internet access and other massive data. Analyze the data to explore the relationship between students' study, life and psychology, dig out the abnormal data of students, feedback the abnormal data, make full use of the data of students' behavior in school to build the digital campus and intelligent campus. So that the level of information on campus can be improved. Based on the subject of ideological and political education of college students of the Beijing Municipal Committee of the Communist Party of China (CPC), this paper builds up an analysis system of student behavior by big data. < Analysis and Application of the data of one Card in Student work in Colleges and Universities from the Perspective of big data. The main contents of this paper are as follows: (1) integrating the historical data of the major business systems of the school, and analyzing the various data in the student's school card, the main contents of the research are as follows: (1) integrating the historical data of the major business systems of the school, And related to the abnormal data processing. (2) study big data framework Hadoop HDFS file system and MapReduce computing model, and build the overall technical framework of college student behavior analysis system based on Hadoop technology. And using the computational model MapReduce to mine the data of college students' behavior. (3) reducing the measuring points of students' behavior data, combing the relationship between different behaviors, and drawing the student's "student portrait" in the school. Clearly describe the situation of students in school, and analyze the relationship among students' learning, living conditions and psychological dynamics. (4) construct the identification model of students with financial difficulties in colleges and universities. By using the concept of membership degree of fuzzy evaluation method combined with the data of student card consumption in big data analysis system and the data of family situation questionnaire, the grade of student membership is determined. The relative size of membership degree is used to determine the poverty grade. (5) to realize the analysis system of student behavior big data, analyze and summarize the law and characteristics of students' behavior, and put forward constructive suggestions for relevant departments to analyze. In order to analyze the characteristics of students' behavior and guide the students' behavior to develop healthily in time. (6) to complete the practical application of big data analysis system in the work of identifying the students with financial difficulties in school families, In order to improve the efficiency of student work and the scientificity and reliability of the result, the way of model identification is used to replace the qualitative analysis of the counselor's experience in the past, and the qualitative analysis of the identification work is changed from the qualitative work to the quantitative one.
【学位授予单位】：华北电力大学(北京)
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP311.13

【参考文献】