基于用户画像的学习资源推送服务模型

A Push Model for Learning Resources Based on User Profile

针对用户在网络学习过程中无法快速精准地找到合适的学习资源的问题,提出了一种两阶段学习资源推送服务模型。第一阶段采用关联规则的推荐服务策略,通过余弦相似度计算用户特征向量与学习资源特征向量的相关度,解决用户学习的冷启动问题;第二阶段采用多维度关联推荐策略,通过分析用户的个人信息和行为数据并结合学习资源的文本信息进行个性推荐,同时动态调整推荐资源内容。试验结果表明该模型能够提高学习资源推荐的准确性和个性化程度,可以有效帮助用户在网络中找到适合的学习资源。该模型可以广泛应用于在线学习领域,具有广阔的推广前景。

This paper proposed a two-stage push model in order to address the problem of users’inability to quickly and accurately find suitable learning resources when studying online.In the first stage,a recommendation strategy,in which the correlation of feature vectors between users and learning resources was calculated by cosine similarity,was designed on the basis of Association Rules and applied to solve the cold-start problem of learning.In the second stage,a recommendation strategy based on multi-dimensional association was utilized to provide individualized recommendation and make dynamic adjustments according to users’personal information and behaviors along with textual learning resources.The Experimental results have demonstrated that the model is conducive to improving the accuracy and personalization of the recommendation,and effectively assisting users in discovering suitable online learning resources.This model can be widely used and popularized in the field of online learning.