HEProOE: A hyperedge enhanced probabilistic optimal estimation method for detecting spatial fuzzy communities.

Area of Science:

• Urban Studies
• Data Science
• Spatial Analysis

Background:

• Human mobility data is crucial for understanding urban spatial structure.
• Current methods often ignore semantic information, fragmenting indivisible regions and causing membership uncertainty.
• Spatial stochasticity in human movement creates ambiguity in fuzzy community boundaries.

Purpose of the Study:

• To propose a novel method, HEProOE, for spatial fuzzy community detection.
• To integrate hyperedges representing indivisible regions (IRs) with probabilistic community membership.
• To enhance community detection by optimizing for both mobility patterns and semantic consistency.

Main Methods:

• Representing indivisible regions (IRs) as hyperedges with probabilistic community membership for each spatial unit.
• Introducing a distance-weighted Jensen-Shannon (JS) divergence metric to quantify semantic consistency within hyperedges.
• Integrating the JS divergence metric as a likelihood component into the mobility-based Probabilistic Optimal Estimation (ProOE) model.

Main Results:

• The HEProOE method effectively integrates human mobility data with semantic information.
• Experimental results show significantly higher semantic consistency in detected spatial fuzzy communities.
• The approach provides a more authentic understanding of urban spatial structures.

Conclusions:

• HEProOE offers a unified framework for spatial fuzzy community detection.
• The method addresses limitations of solely relying on mobility data by incorporating semantic consistency.
• This approach enhances the accuracy and interpretability of urban spatial community analysis.