A citation recommendation method based on dual-channel heterogeneous hypergraph neural networks

LI Ruihong, LI Xiao Hong, YAO Jin, WANG Shanshan

Abstract: Addressing the issue that existing citation recommendation methods primarily focus onmodeling binary relationships using graph structures, and lack sufficient representation of the diversityand variety of node types and interaction relationships, a citation recommendation method based on dualchannel heterogeneous hypergraph neural networks is proposed, Firstly, a heterogeneous graph is constructed, convolutional neural networks (CNNs) and 'Transformers are utilized to encode the local andglobal semantic features of each node in the heterogeneous graph, respectively, obtaining structural representations of the target node on the heterogeneous graph channel. Secondly, multiple types of hyper.edges are designed to expand heterogeneous data information, Thirdly, a hypergraph is used to encodeinteractions between nodes, and a hypergraph neural network is employed to capture potential complexhigh-order semantic relationships in the hypergraph, obtaining semantic representations of the targetnode on the hypergraph channel. Finally, information from the two channels is aggregated to obtain thefinal semantic representation of the target node. The correlation between the target paper node and candidate paper nodes is calculated to generate a citation recommendation list. Experimental results on theDBLP and PubMed datasets demonstrate that the proposed method can effectively improve the quality ofcitation recommendations and achieve better recommendation outcomes.