Force-Directed Graph Layouts Revisited: A New Force Based on the T-Distribution.

Area of Science:

• Computer Science
• Graph Theory
• Data Visualization

Background:

• Force-directed graph layout algorithms are crucial for visualizing complex networks.
• Existing methods often struggle with preserving local neighborhood structures or computational efficiency.
• Adaptability in force parameters is limited in current approaches.

Purpose of the Study:

• To propose a novel force-directed placement model, t-FDP, for improved graph layout.
• To enhance neighborhood preservation and maintain low stress errors in graph visualizations.
• To develop an efficient implementation for real-time parameter tuning on complex graphs.

Main Methods:

• Introduction of a bounded short-range force (t-force) based on Student's t-distribution.
• Flexible formulation allowing separate adaptation of short- and long-range effects.
• Efficient implementation utilizing Fast Fourier Transform (FFT) for accelerated computation.

Main Results:

• t-FDP demonstrates superior neighborhood preservation compared to current methods.
• The model achieves low stress errors, indicating high-quality layouts.
• FFT-based implementation offers significant speedups: one order of magnitude faster on CPU and two orders faster on GPU.

Conclusions:

• The t-FDP model offers an effective and efficient approach to force-directed graph layout.
• Its speed and adaptability facilitate real-time parameter tuning for interactive exploration of complex graphs.
• The method provides a valuable advancement for data visualization and network analysis.