To address the current issue of internet information overload,recommendation systems have been widely used in e-commerce,news and music websites,and other fields.The goal of rec-ommendation algorithms is to explore users'potential interests,provide personalized information push,and ultimately solve the problem of information overload.Therefore,it is particularly impor-tant to choose a suitable recommendation algorithm to solve the classification problem for users,and it is also an important method to solve the field of data mining.However,the current recommenda-tion model lacks effective utilization of heterogeneous data from multiple sources,and there is a problem of information loss in the process of aggregating semantic information.To address the above issues,this article proposes a multi-dimensional semantic fusion recommendation algorithm model based on heterogeneous information networks.Firstly,the complex semantic structure within the rec-ommendation task is described through meta paths and heterogeneous graphs.Then,the neighbor-hoods guided by the meta paths are divided,and multi-scale semantic information is captured through multi-layer neighborhood interaction.Finally,multi-scale semantic information fusion is guided in low and high order dimensions.The experimental results show that this method has high accuracy.
关键词
人工智能技术/异质信息网络/元路径/语义融合
Key words
artificial intelligence technology/heterogeneous information network/metapath/seman-tic fusion
维普
万方数据