Application of PSO-integrated K-means algorithm in resident digital portrait classification.

Area of Science:

• Computer Science
• Data Science
• Public Administration

Background:

• Digital governance relies on accurate resident profiling for administrative efficiency.
• Traditional K-means clustering faces limitations with high-dimensional and complex resident data.

Purpose of the Study:

• To introduce a novel hybrid algorithm, Particle Swarm Optimization-K-means (PSO-KM), for improved resident data clustering.
• To enhance the accuracy and computational efficiency of digital profiling in local-level governance.

Main Methods:

• Integration of Particle Swarm Optimization (PSO) for global optimization with K-means for iterative refinement.
• Dynamic updating of cluster centroids using PSO-KM on comprehensive resident data from 2023.
• Comparative analysis against conventional (e.g., GA-K-means) and advanced (e.g., DBSCAN) clustering methods.

Main Results:

• PSO-KM demonstrated superior clustering performance with a silhouette score of 0.752 ± 0.021 and inter-cluster distance of 1.493 ± 0.036.
• Behavioral data exhibited the highest classification performance (silhouette value 0.184), indicating the significance of dynamic traits.
• Segmentation analysis revealed distinct dominant features across income levels: demographic (low-income), behavioral (middle-income), and social network (high-income).

Conclusions:

• The PSO-KM algorithm offers a significant advancement in resident profiling for digital governance.
• Insights derived from PSO-KM segmentation can inform targeted community management strategies.
• The study highlights the potential of advanced clustering techniques to refine grassroots digital governance practices.