Citation Classification Based on Formal Concept-Enhanced Graph Convolutional Network

Citation networks, as an important carrier of academic communication and knowledge dissemination, face with the problem of insufficient data mining and analysis due to its large scale. This paper proposes a method based on Formal Concept Analysis (FCA) to optimize the extraction of structural features of citation networks. By identifying the concept hierarchy and association relationships in the network, FCA can effectively capture the complex structural information within the citation network, thereby improving the quality of feature representation and further data mining. On this basis, this paper uses Graph Convolutional Network (GCN) to model the optimized structural features and builds a deep learning model for node classification. This model makes full use of local neighborhood and global topological information in the citation network to improve the accuracy and robustness of downstream tasks. Experimental results demonstrate that the proposed method achieves superior performance in the node classification task, improving accuracy by approximately three percentage points compared to traditional methods. This work addresses the data mining problem caused by the huge amount of citation network data. The method in this article provides a new technical solution for in-depth analysis and application of citation networks, and has broad application prospects.