Hypergraph-Clustering Method Based on an Improved Apriori Algorithm

With the complexity and variability of data structures dimensions, traditional clustering algorithms face various challenges. The integration network science has become a popular field exploration. One main challenges is how to handle large-scale complex high-dimensional effectively. Hypergraphs can accurately represent multidimensional heterogeneous data, making them important for improving performance. In this paper, we propose hypergraph-clustering method dubbed “high-dimensional method” based on hypergraph partitioning using an improved Apriori algorithm (HDHPA). First, constructs association rule algorithm, where frequent itemsets existing in are treated as hyperedges. Then, different mined parallel obtain hyperedges with corresponding ranks, avoiding generation redundant rules mining efficiency. Next, use dense subgraph partition (DSP) divide into multiple subclusters. Finally, merge subclusters through sub-hypergraphs results. advantage lies its model discretize between space, which further enhances effectiveness accuracy clustering. We comprehensively compare proposed HDHPA several advanced methods seven types datasets then their running times. results show that evaluation index values generally superior all other methods. maximum ARI value reach 0.834, increase 42%, average time lower than All all, exhibits excellent comparable performance real networks. research paper provide effective solution processing analyzing also conducive broadening application range techniques.