Author(s)

Huan Lian

Corresponding Author

Huan Lian

Abstract

Personalized push of multimedia Japanese education materials in Colleges and universities can improve the efficiency and adaptive matching level of Japanese teaching materials. This paper proposes a personalized push algorithm of multimedia Japanese education materials in Colleges and Universities Based on multi task deep learning. This paper constructs a deep learning model of multimedia Japanese education materials under the multi view fusion mechanism. By filtering and recommending the content, it carries out adaptive knowledge mapping analysis and deep learning on the summary, title, semantic knowledge, etc. of the learning materials. According to the dependency of semantic knowledge, it establishes a multi task deep learning model of multimedia education materials in Colleges and universities. In the process of deep learning, it mines the deep semantic correlation feature quantity, and realizes the personalized recommendation of multimedia Japanese education materials according to the clustering results of the feature quantity. The simulation results show that this method can improve the accuracy of recommendation, effectively use the automatic classification results in different semantic states to achieve data recommendation, and is superior to various views and models, can better achieve the automatic classification and category recommendation of Japanese teaching resources, effectively use the multi view knowledge and semantic information, and improve the efficiency of multimedia Japanese education materials.

Keywords

Multi Task Deep Learning; Multimedia Japanese Education Materials in Colleges and Universities; Personalized Push; Context; Semantics