<scp>PCA</scp> ‐ <scp>XGBoost</scp> Combined With K‐Anonymity and Random Forest Privacy‐Protecting Clustering Algorithm

ABSTRACT In response to the actual needs of privacy protection and cluster analysis of agricultural product traceability data, this work innovatively proposes a privacy‐protecting clustering algorithm combining PCA, XGBoost, K‐anonymity and Random Forest. Experiments were conducted on a constructed rice traceability dataset, with evaluations from three aspects: clustering accuracy, privacy protection effect, and computational efficiency. Results show that the algorithm achieves 0.836 in Adjusted Rand Index and 0.782 in Silhouette Coefficient, representing significant improvements over traditional methods. Meanwhile, the privacy leakage risk was reduced to 6.7%, balancing privacy protection and clustering performance. This work provides a reliable technical solution for agricultural product quality supervision and supply chain optimization.