A cross-course video subgraph recommendation method based on map association

Abstract

The invention discloses a cross-course video sub-graph recommendation method based on map association, which comprises the following steps: 1) extracting video learning behavior characteristics by using learner's network learning log based on implicit feedback; 2) obtain that videoSeeds of a seed video set; 3) Constructing knowledge points across courses Video association map; 4) Making use of cross-curriculum knowledge points Video Correlation Map Calculation Course Video Correlation Degree; 5) By using the implicit characteristics of user behavior, take a seed video set, Combining with course video relevance degree, this method provides learners with cross-course video subgraph containing knowledge relevance.This method can synthetically consider the characteristics of users' implicit learning and the knowledge relevance between course videos, and provide online learning platform learners with course video recommendation satisfying their learning preferences and considering cross-course knowledge relevance.

Description

technical field [0001] The invention relates to a method for recommending course video subgraphs, in particular to a method for recommending cross-course video subgraphs based on map association. Background technique [0002] Different from traditional education, online education has transformed from the traditional teacher-led classroom to the learner's autonomous learning model. At the same time, online education provides learners with richer learning resources, such as video and audio, PPT, exercises, etc., and learners can choose from them according to their own needs. However, in the face of massive learning resources, learners are prone to cognitive overload [1] At the same time, due to the large differences in the educational background of the learners and the uneven learning ability, it is also easy to cause the unreasonable arrangement of the learners' learning progress and affect the learning effect. The recommendation system is an effective way to solve such pro...

AI Technical Summary

Problems solved by technology

However, there is still no mature general solution in the field of learning resource recommendation because of the uniqueness of the online learning platform itself: First, the online learning platform has a large number of learning resources, but often lacks learners' explicit learning resources such as course videos. Rating data, unlike Douban and other platforms, users will have star ratings for movies, books, etc. The level of ratings can accurately reflect the user's preference for the rated items, so the recommendation system can directly use rating data and use collaborative Filtering and other methods are used to recommend users; second, to recommend learning resources requires a certain amount of domain knowledge. The courses of different majors vary greatly, and the knowledge of different courses in the same major is related. Content-based learning resources Recommendations are obviously much more complicated

However, the existing learning resource recommendation algorithms often ignore the correlation between learning resources (such as course videos). This correlation exists not only within the course, but also among multiple courses of similar majors. The association relationship between videos is more limited within a course, and the recommendation results are mainly in the form of resource lists, which will weaken the knowledge association between videos

Images