Research And Implementation Of Personalized Question Recommendation System Based On Knowledge Proficiency

In the highly developed Internet industry today,humans have stepped into the big data.The application of big data is increasingly universal,the explosion of information and the growing acceptance of big data technology have combined to foster continuously development of personalized recommendation technology.Online recommender systems in various fields have made people’s daily life more convenient,helped people filter information,and improved user satisfaction.The same is true of the development of education.How to make teachers and students use the valuable resources quickly in the massive e-learning resources has become one of the research focuses.This thesis,in conjunction with pertinent educational theory,makes a further research and analysis of the present circumstances in the realm of e-learning and personalized recommendation both domestically and internationally.The study revealed that,although many service platforms apply recommendation algorithms,the educational sector in the education industry has been unable to satisfy the current market requirement for a single recommendation algorithm’s implementation.For instance,the examination question recommendation algorithm based on collaborative filtering ignores the information of students’ knowledge mastery level,and only relies on similar ratings or the inherent attributes of resources to recommend,which is easy to give too simple or difficult exercises,resulting in unreasonable recommendation results.The exercise recommendation method based on knowledge tracing mainly models and predicts the individual knowledge status of students,ignoring the similarity between student groups.In addition,the existing problem recommendation strategies which are mainly formulated by authoritative experts or sorted according to the grade of questions resources.Therefore,it is difficult to timely correct the strategic orientation according to the change of students’ cognitive level.This thesis presents a research and implementation of a personalized test questions recommendation system based on knowledge proficiency,taking into account this.This study synthetic analyzes the recommendation algorithm,expounds the principles,advantages and disadvantages of several typical recommendation algorithms.By combining different algorithms,completes the advantages and disadvantages of the algorithm and functional complementarity,so that improve the efficiency and quality of recommendation.At the same time,combining with the relevant educational theories to establish test question recommendation strategy,devised the individuation test questions recommendation system that conforms to the educational theories.Specifically,based on the user’s historical answer data,the deep knowledge tracking model was used to model the student’s knowledge level,and then the Canopy-K-means clustering algorithm was introduced to diminish the complexity of user similarity computation and elevate the calculation efficiency of the overall system.Finally,the application searches for students with similar knowledge level to the target user from the current user cluster according to the user’s collaborative filtering algorithm,and predicts the student’s grade based on the student’s own knowledge level and the knowledge level of similar student teams.In the question recommendation strategy,personalized recommendation for target students,respectively recommending questions for expansion and improvement and questions for review and consolidation.The former relies on the program teaching theory and humanistic learning theory,and selects suitable difficulty test questions to recommend for students to expand and improve.According to the forgetting curve,the latter recommends the answers with higher forgetting degree to students for review and consolidation,and finally produces the effect of personalized recommendation.This thesis,based on the analysis of system requirements,has designed and implemented an individuation test questions recommendation system based on knowledge proficiency.The recommendation system mainly is divided into three modules,including administrator terminal,teacher terminal and student terminal.The student terminal mainly provides online examination,exercise recommendation,error book other services to students,while the teacher terminal mainly provides the question bank management,test paper management,and other tasks.The administrator terminal mainly provides teacher information management and student information management functions.This system uses Java language development,the front and back end are built by Bootstrap and SSM framework respectively,and pays attention to the impact on system fluency and data processing in algorithm and technology selection.In conclusion,this thesis actualizes an individuation questions recommendation system based on knowledge proficiency through a large number of theoretical research,carrying out corresponding comparative experiments and related tests of the recommendation system.This research can enable the system to recommend test questions that are in line with their knowledge proficiency for users,which can effectively improve the quality of learning.