基于协同过滤的高校数字化教材资源推荐方法

The Recommendation Method of University Digital Textbook Resources Based on Collaborative Filtering

在开展学习资源推荐的过程中,由于用户行为的复杂性和资源的多样性,导致推荐结果的可靠性偏低,为此提出基于协同过滤的高校数字化教材资源推荐方法。以用户的基本特征参数为基础,对用户进行聚类处理。在资源推荐阶段,结合相同聚类其他用户的意见,向目标用户推荐教材资源,形成资源与用户之间的邻域关系后,估计用户对教材资源的具体评分,按照评分由高到低的顺序进行相应的资源推荐。测试结果表明,该方法的命中率和归一化折扣累积增益优于对照组,具有较高的可靠性。

In the process of recommending learning resources,due to the complexity of user behavior and the diversity of resources,the reliability of recommendation results is low.Therefore,a collaborative filtering based recommendation method for digital textbook resources in universities is proposed.Cluster users based on their basic feature parameters.In the resource recommendation stage,combined with the opinions of other users in the same cluster,recommend textbook resources to the target user,form a neighborhood relationship between the resources and the user,estimate the specific rating of the user on the textbook resources,and recommend corresponding resources in the order of rating from high to low.The test results show that the hit rate and normalized discount cumulative gain of this method are superior to the control group,and have high reliability.