The Research Of Personalized Learning Resource Recommendation Based On Learning Diagnosis

With the Internet technology developing in the field of education applications,more and more people focus on personalized learning.The rapidly growing learning resources,including online and offline resources,makes learners encounter "information overload" and "information maze" confusion.In the face of massive learning resources,it is of great challenge to solve the problems that how to dig and recommend appropriate learning resources in limited time in the field of educational information.Compared with e-commerce,the information dig and recommendation in the field of education have their own characteristics,that is,learning target characteristics.The user must consider the learning target as well as the interested resources.Interest preference is not the only indicator of resource recommendation.The user's learning goal is also an important factor in the recommended process.In this paper,the personalized recommendation of learning resources is proposed based on learning diagnosis.Firstly,the connection diagram of knowledge points is built based on the ontology.The knowledge points which the students master are determined according the test knowledge point matrix and test scores matrix.On account of the points which are failed to master,specialized knowledge points sequence is generated by topology sorting method.Then,we classify the multi-dimensional behavioral data,analysis and establish its relationship with the learning effect,thus discovering students' learning habits and interest.Thirdly,according the learning target and learning preference,the personalized recommendation of learning resource is built and the collaborative filtering recommendation algorithm is designed.To avoid the knowledge blank,the parallel knowledge points are recommended after the personalized knowledge points in the topological sequences having been mastered.At last,we conduct the experiments on the data generated by the learning platform.Experiments results show that the recommended method of personalized learning resources can improve the accuracy and recommendation effect.