Research And Application Of Undergraduates’ Learning Behavior Analysis

This dissertation has done a lot of research on undergraduates’ learning behavior analysis.Based on the undergraduates’ online data,this dissertation puts forward the concept of website contribution to professional learning and gives the calculation method.Based on this,it constructs a model of undergraduates’ learning behavior analysis and implements the academic warning system.Under the background of the full implementation of the credit system,the system promptly reminds undergraduates to maintain a good learning status and helps them graduate smoothly.The main research contents are as follows:(1)Propose the concept of website contribution to professional learning,and give the calculation method.Firstly,this dissertation establishes a professional thesaurus in combination with the curriculum system in the training program,and uses the language model algorithm to expand the professional thesaurus.Secondly,using a variety of technologies to obtain and process undergraduates’ online content,and using feature words to express undergraduates’ online content.Finally,the website contribution degree algorithm based on keyword matching(WCDA-KM)is proposed for calculating the degree of contribution of the website.The higher the degree of relevance between undergraduates’ visit content and the major they studied,the higher the degree of contribution of the website,thus discovering the undergraduates’ motivation for surfing and determining the learning status of undergraduates.(2)Study the key attribute of online data which affects the performance of undergraduates and formulate dynamic warning rules for the key attribute.In this dissertation,the improved Naive Bayes algorithm(NB)is used to analyze undergraduates’ online data and achievement data.Since the NB algorithm assumes that the properties are independent of each other and can not satisfy the characteristics of the attributes in the online data,this paper improves the NB algorithm by introducing the feature weighted idea and using the improved flower pollination algorithm(IFPA)to search the global optimal attribute weights.The Naive Bayesian classifier based on improved flower pollination algorithm(NBC-IFPA)is proposed.The NBC-IFPA algorithm is used to train online data and achievement data.The optimal attribute weights obtained are used to determine the key attribute that affects the learning status of undergraduates,and dynamic warning rules are determined for the key attribute and used for system implementation.(3)Implement the academic warning system.Based on the degree of contribution of the website,the academic warning system is implemented,including dynamic warning and static warning.The dynamic warning is based on dynamic warning rules,which dynamically tracks the learning status of undergraduates and gives warning to undergraduates who meet the warning rules.The static warning determines the academic warning rules through in-depth analysis of educational data and the academic management system which gives prompt warnings and help to undergraduates who deviate from normal learning status.