Variable Selection for High-dimensional Nodal Attributes in Social Networks with Degree Heterogeneity

Area of Science:

• Network analysis
• Statistical modeling
• Machine learning

Background:

• Network analysis often faces challenges with ultrahigh-dimensional data.
• Understanding homophily and degree heterogeneity is crucial for accurate network modeling.

Purpose of the Study:

• To propose a Bayesian method for nodal feature selection in dense and sparse networks.
• To develop a computationally efficient working model for large sparse networks.

Main Methods:

• A Bayesian approach using Gibbs sampling for feature selection.
• Development of a working model for computational efficiency in large networks.
• Asymptotic model selection consistency analysis.

Main Results:

• The proposed Bayesian method effectively selects nodal features.
• Model selection consistency is proven, even with exponentially growing dimensions.
• The working model alleviates computational burden for large sparse networks.

Conclusions:

• The developed Bayesian method provides a robust framework for feature selection in complex networks.
• The approach is validated through simulations and real-world data analysis.
• This work advances network analysis techniques for high-dimensional data.