【摘要】 为提高推荐系统的性能,将评分预测任务和排序任务在共享用户和item表示的基础上集成在一个多任务学习框架中,在训练过程中同时优化两个任务的参数集。为进一步提高泛化效果,将用户的决策过程分为两个阶段,即用户选择一个item进行交互(排序任务),再对其评分(评分预测任务)。在此基础上,提出一个可融合不同底层算法的通用多任务框架,在两个数据集上通过实验对其进行评估,实验结果表明,其优于现有的最先进的方法。更多还原

【Abstract】 To improve the performance of the recommendation system,the rating prediction task and the ranking task were integrated in a multi-task learning framework based on the shared user and item representation,and the parameter sets of the two tasks were simultaneously optimized during the training process.To further improve the generalization effects,the user’s decision process was divided into two phases,the user first selected an item to interact(ranking task),and then rated it(rating prediction task),on this basis,ageneral multi-task framework that could fuse different underlying algorithms was proposed and it was evaluated by experiments on two data sets.The results show that it is superior to the existing state-of-the-art methods.更多还原