Feasibility of the development of short videos in vocational education based on the context of deep learning

Abstract

The development of information technology has brought new opportunities for the development of vocational education teaching, and short video, as a product of information technology, breaks through the limitations brought by the traditional teaching mode and provides a new thinking direction for the transformation of the teaching mode. This paper proposes a recommendation system for short videos based on deep learning, where the scene and behavioral features of the target short video are first extracted to obtain the joint features. Then, user and video modeling is realized using deep learning technology. Finally, the similarity and preference are calculated by combining the joint features of the target short video and the user short video preference features. The TOP-N with the highest user preference is obtained according to the actual demand. The recommendation list is generated to complete the intelligent recommendation of short videos based on deep learning. The experimental analysis found that the recommendation system has high performance, with a maximum recommendation accuracy of 98.2%, and is capable of fully comprehending the students’ demand preferences for short videos. The use of short videos in course teaching can make a great positive change in students’ learning attitudes, and the short video recommendation algorithm based on deep learning allows students to more accurately find short videos suitable for their learning conditions, which further improves their learning efficiency.