Analysis of airport passenger behavior patterns based on clustering algorithms

Analysis of airport passenger behavior patterns based on clustering algorithms aims to reveal the potential laws of passenger behavior through data mining technology and provide decision support for airport management optimization. The study constructs a high-quality feature set through systematic data preprocessing, including cleaning, feature selection, standardization and dimensionality reduction. Comprehensively comparing K-means, DBSCAN and hierarchical clustering algorithms, the optimized K-means algorithm stands out with its high efficiency and stability, and divides passengers into four groups: efficient business, leisure shopping, short stay and long waiting. The groups have significant differences in check-in time, security waiting time, consumption amount and other characteristics. The clustering results are verified by silhouette coefficient and cross-validation, which proves the reliability and application value of the model. The research results provide a scientific basis for airport resource allocation, service personalization design and operational efficiency improvement, and have important theoretical significance and practical value. © 2026 SPIE.