Advanced Network Sampling with Heterogeneous Multiple Chains.

Area of Science:

• Computer Science
• Network Science
• Data Science

Background:

• Large-scale networks like the Internet of Things and social networks are growing exponentially.
• Studying these networks is challenging due to their size and limited data access.
• Existing sampling methods struggle with the scale and complexity of modern networks.

Purpose of the Study:

• To propose a novel and effective sampling method for large-scale networks.
• To address the limitations of current random-walk-based sampling techniques.
• To enable efficient exploration and analysis of vast network data.

Main Methods:

• Developed a novel sampling approach utilizing multiple heterogeneous Markov chains.
• Adjusted random-walk characteristics to efficiently explore the target network space.
• Conducted experiments on both synthetic and real-world large network datasets.

Main Results:

• The proposed method achieves better unbiased sampling results.
• Demonstrated reduced asymptotic variance compared to existing methods.
• Achieved efficient network exploration within reasonable execution times.

Conclusions:

• The novel Markov chain-based sampling method is effective for large-scale networks.
• This approach outperforms current random-walk-based sampling techniques.
• Offers a promising solution for analyzing complex and massive network structures.