Characterizing spreading dynamics of subsampled systems with nonstationary external input.

Area of Science:

• Complex systems dynamics
• Statistical modeling
• Computational neuroscience
• Epidemiology

Background:

• Many systems exhibit propagation dynamics, often modeled using autoregressive models.
• Estimating parameters in these models is challenging due to subsampling and time-dependent parameters.
• Existing methods can lead to incorrect estimates under experimental constraints.

Purpose of the Study:

• To develop an analytical method to overcome subsampling bias in autoregressive models.
• To accurately estimate propagation rates in systems with nonstationary external input.
• To provide a robust approach applicable to real-world experimental data.

Main Methods:

• Analytical derivation to correct for subsampling bias.
• Application of the method to autoregressive models with nonstationary input.
• Validation using simulated data and real-world datasets.

Main Results:

• Successfully demonstrated how to analytically overcome subsampling bias.
• Accurate estimation of propagation rates achieved even with partial system observation.
• Method validated on neural spike data and infectious disease spread (norovirus, measles).

Conclusions:

• The proposed analytical approach effectively corrects for subsampling bias.
• This method enhances the reliability of parameter estimation in propagation models.
• Applicable to diverse fields including neuroscience and public health for improved data analysis.