Monte Carlo Expectation Maximization addresses missing data in non-stationary multivariate time series, highlighting implications for psychiatric research.
Key Points
The proposed MCEM-SSM effectively addresses missing data challenges in non-stationary multivariate time series.
Simulations demonstrate MCEM-SSM's advantages over conventional imputation methods for non-stationary data.
This approach was validated through real-world data analysis of bipolar and schizophrenia patients.
Findings suggest a link between digital social connectivity and negative mood, significant for psychiatric illness.