Clustering Identification and Association Rule Mining of Urban Expressway Cross-Section Cases

Urban expressway renewal cases are characterized by scattered information and difficulty in extracting engineering regularities.To address these issues, this study conducts clustering and association rule mining on cross-section cases and validates lane configuration patterns using expressway traffic data.The results show that the cross-section cases can be classified into five typical patterns, with an overall silhouette coefficient of 0.534.Relatively stable associations are identified among ground-level frontage roads, non-motorized vehicle lanes, elevated mainlines, and cross-section widths of 25~35 m.In addition, as traffic volume increases, the proportions of 8-lane and 10-lane schemes increase significantly.These findings provide a reference for the systematic compilation of expressway renewal cases and for preliminary scheme analysis.