【摘要】 传统的推荐模型主要是基于用户或者基于项目层面进行建模分析，而未考虑过用户与项目之间存在的协作信号。针对上述问题，提出结合图神经网络的推荐模型，将用户与物品数据构造为图结构，将图结构输入神经网络中挖掘用户与物品之间的协作信号，使得用户和物品得到更具体的特征表示。实验结果表明，这一改进在一定程度上提高了模型的准确度，使得模型推荐效果得到提升。更多还原

【Abstract】 The traditional recommendation model is primarily based on the user or project level for modeling and analysis, but has never taken the collaboration signal between the user and the project into consideration. Aimed at the above problem, a recommendation model based on graph neural network is proposed. The user and item data were constructed into a graph structure, which was input into the neural network to exploit the collaboration signal between the user and the item, so that the user and the item could get more specific feature representations. The experimental results indicate that the accuracy of the model is improved to a certain extent, and the model recommendation effect is improved.更多还原