Accounting for edge uncertainty in stochastic actor-oriented models for dynamic network analysis.

Area of Science:

• Network analysis
• Statistical modeling
• Computational neuroscience

Background:

• Stochastic Actor-Oriented Models (SAOMs) are standard for analyzing social network dynamics.
• The assumption of error-free network data in SAOMs is often unrealistic.
• Real-world networks frequently contain false positive and false negative edges.

Purpose of the Study:

• To develop an extension of SAOMs that accounts for noise in observed network data.
• To improve the accuracy of network change estimation in the presence of measurement error.
• To apply the novel method to functional brain network analysis.

Main Methods:

• A hidden Markov model (HMM) framework integrating a latent Markov process for true networks and a measurement model for observed networks.
• An expectation-maximization algorithm for parameter estimation.
• Particle filtering and the missing information principle to handle large state spaces.

Main Results:

• The proposed HMM-SAOM extension demonstrated improved estimation accuracy compared to standard SAOMs on noisy network data.
• Simulation studies confirmed the enhanced performance in the presence of measurement error.
• Application to electroencephalogram (EEG) data revealed larger effect sizes in functional brain networks.

Conclusions:

• The HMM-SAOM extension provides a more robust approach for analyzing dynamic networks with noisy observations.
• This method offers significant advantages over standard SAOMs, particularly in fields like neuroscience.
• Accurate modeling of network dynamics with measurement error is crucial for reliable scientific inference.