基于差异化资源分配的二部图推荐算法

Bipartite Graph Recommendation Algorithm Based on Differentiated Resource Allocation

传统二部图推荐算法存在着初始资源设置不合理,以及仅仅依靠项目度和用户度来调整资源分配系数的问题。因此提出一种基于差异化资源分配的二部图推荐算法,新的算法利用评分规范化和最大最小值的方法对项目初始资源进行了修正,在此基础上引用艾宾浩斯遗忘函数来量化用户"兴趣偏移"所带来的影响;再利用用户评分相似性函数和用户偏好函数对资源分配系数进行了差异化设置,使资源流转变得更加合理。经过实验验证,新提出的算法在推荐准确度及多样性上都有所提升。

The traditional bipartite graph recommendation algorithm has the problems that the initial resource setting is unreasonable,and the resource allocation coefficient is adjusted only by the project degree and the user degree.Therefore,a bipartite graph recommendation algorithm based on differential resource allocation is proposed.The new algorithm uses the scoring normalization and maximum and minimum methods to correct the initial resources of the project.On this basis,the Ebbinghaus forgetting function is used to quantify the user.The impact of"interest offset";the user score similarity function and the user preference function are used to differentiate the resource allocation coefficients,making the resource flow more reasonable.After experimental verification,the proposed algorithm has improved in recommendation accuracy and diversity.