Research On Curriculum Relevance And Comprehensive Quality Evaluation Of Students Based On Students' Learning Outcomes

Due to the continuous development of informatization construction in colleges,colleges have stored massive amounts of student data.The database in the educational management system of colleges is an important data source of student data.However,this data is currently only used for simple query and statistics,in spite of tremendous information hidden in it.Therefore,exploring the intrinsic value of educational big data through data mining technology plays a crucial role in the innovation reform of university management and personnel training.This thesis conducted research on curriculum relevance and students' comprehensive quality evaluation according on the original data of the communication engineering students enrolled between 2009 and 2012 in the college's educational management system.The data consists of their course grades,their awards for extracurricular competitions and their graduation design.Firstly,focusing on the grades and courses,the adaptive multi-minimum support association rule algorithm is used to deeply explore the relevance of the courses.Secondly,focusing on students,principal component analysis and the SOM neural network are applied to deeply analyze the comprehensive quality of the individuals and various groups of students.Finally,based on the results of mining and analysis,this thesis provides useful information for the optimization of the curriculum system and the promotion of students'comprehensive and balanced development.The main contributions are as follows:(1)Firstly,based on the whole picture of the distribution of the course grades,and combined with statistics and discretization requirements,a method was adopted that divides grades by means of standard deviation with average grades as the center.Secondly,according to the colleges' grading rules,the quantitative treatment of the awards for extracurricular competitions and the graduation design of students was completed.(2)As for the classic association rule algorithm,it is difficult to mine the relevance of small support events.Featured with a different frequency of the grade distribution based on course grades,the statistical fitting technique is applied to determine the support and the confidence threshold.And the Confidence-lift model is introduced to screen out valuable rules.Therefore,the adaptive multi-minimum support association rule algorithm is proposed to analyze curriculum relevance,the relevance of different courses in the communication engineering major is obtained.By comparing the experimental results of the traditional Apriori algorithm,the effectiveness of the improved association rule algorithm is verified.(3)Based on the scores of the students'internal and external grades after the quantitative processing,the principal component analysis method was applied to construct the comprehensive quality evaluation index of college students in nine dimensions.At the same time,combined with the radar map to make a comprehensive and multi-dimensional evaluation of the individual students,and the validity of the individual's comprehensive ranking is proved by comparing the student's research and retention status.According to the normalized values of each dimensional index,the classification of student groups is completed by the SOM neural network.And the comprehensive quality development level of each group of college students is compared and analyzed.