Research On The Feature Model Of Digital Education Resources Integrating Metadata And Social Annotation

In the process of promoting educational informatization,it gradually appears that digital education resources are overloaded,resulting in resource users in a state of "information" loss,unable to quickly find the resources to meet their needs,and most of the high-quality resources are also buried because of low utilization rate.Recommendation technology can accurately recommend personalized learning resources to users by mining and analyzing valuable information in the data of resource attributes and characteristics.Among them,how to effectively acquire and analyze attribute feature data sets and representation resources,and establish resource feature model accordingly,which is the research focus of improving recommendation service quality.Existing researches mostly use expert metadata to represent digital education resources.The representation dimension is relatively single,and it is difficult to truly evaluate resources from the perspective of users.In addition,the operation-oriented resource representation is less,and there is no comprehensive and detailed model definition,model building strategy and model application description.This research will focus on the fusion of metadata and social tag to mine and describe the attribute characteristics of more dimensions of resources,and use social networks analysis to make up for the lack of semantic control of social tagging,build an M-S-S resource model and simulate the application.The main research includes:(1)Realize the characteristic model of digital education resources.Select the LOM and CELTS-42 metadata specifications to determine the resource attribute framework,and directly and accurately describe the attributes that can be simply defined in the framework;the attributes that are not easy to be defined in the framework are described by means of social annotation.The social tagging system based on WeChat applet is designed,developed and applied to collect user tags.The centrality,relevance and aggregation of high-frequency tags are analyzed by using social network analysis method to obtain the hierarchical structure of resource tags.(2)Verify the application performance of the model.Design therecommendation algorithm of mixed resource characteristic model and item collaborative filtering;use simulation and tensor expansion techniques,and Movielens is used as data set to verify the actual recommendation performance of the algorithm.The experimental results show that compared with the item-based collaborative filtering recommendation algorithm,the mixed algorithm can alleviate the problem of data sparsity and cold start,with higher recommendation accuracy,and the possibility of resources being found and recommended from different feature dimensions is increased.(3)Summarize the research and application trend of digital education resource model.With the continuous expansion of modeling data sources,as well as the continuous improvement and update of feature disclosure methods such as metadata and social annotation,the mining and description of resource characteristics will tend to be multi-dimensional.The model constructed based on the above can be used to improve the traditional recommendation algorithm,multi-dimensional clustering of digital education resources,and establish a mapping relationship with the user feature model,etc.