Analysis And Research Of Student Behavior Based On Data Mining Technology

Big data has had a huge impact on people’s lives,work and study.As a result,educational informatization has developed rapidly.The digital construction of the university’s campus has been continuously improved.Various applications such as the all-in-one card platform,educational administration system and library management system built in universities have gradually matured.These systems generate a large amount of student behavior data every day.These data resources provide strong data support for the hidden information mining of university student behavior data.At present,universities urgently need to use technical means to help education managers optimize their education management plans for students.The graduation thesis designs and implements a student behavior analysis and early warning system.In order to make the system more efficient and accurate,statistical analysis and data mining technology are used to conduct comprehensive profile analysis and data mining analysis on student behavior data.assisting the university to explore the potential relationship between student behavior and student life and learning.details as follows:1.The HDWA-Kmeans algorithm is designed.The function of this algorithm is to assist the identification of poor students in schools.The algorithm is improved in two aspects on the basis of the classic clustering algorithm K-means algorithm.The clustering effect is better.There are mainly two improvements:(1)Based on the improvement of high-density clustering,it is possible to select a more effective initial cluster center;(2)By adding a threshold coefficient to change the way of recalculating the centroid,the use of a weighted average can ensure the clustering.Such algorithms must be able to converge.In the process of clustering,the differences between data objects can be quantified.The clustering speed is faster.In order to verify the efficiency and accuracy of the improved algorithm HDWA-Kmeans,a clustering experiment was compared based on the consumption behavior data of the students of Minzu University of China.2.The MCDM-Apriori algorithm is designed.The function of the algorithm is to help early warning of students’ possible missed subjects.Based on time and space considerations,the classic association analysis algorithm Apriori algorithm is improved.The improved MCDM-Apriori algorithm is obtained.There are mainly two improvements:(1)The improvement of the Apriori algorithm based on matrix compression.By optimizing the pruning strategy in the algorithm calculation process,it can effectively reduce the number of intermediate candidate sets and reduce the space consumption of the computer;(2)Improved Apriori algorithm based on divide and conquer and merge.By dividing a large database into multiple small databases that are independent and unrelated.Parallel calculation of these small databases respectively improves the utilization rate of the computer and reduces the computing bottleneck of the algorithm.In order to verify the efficiency of the improved algorithm MCDM-Apriori algorithm,an association analysis experiment is carried out based on the study behavior data of the students of Minzu University of China.3.Designed and implemented a student behavior analysis and early warning system,specifically:(1)Performed functional and non-functional analysis of system requirements,designed system architecture,system functions,and system database,etc.;(2)Detailed design and implementation of data mining based on spark platform,user login module,student comprehensive portrait module,poor student assistance identification module,system management module,and comprehensive student early warning module.To help student work managers make more accurate decisions in the identification of poor students and academic early warning.