Extended Data Fig. 2: Convergence Diagnostics for Random-Walk HMC Fit. | Nature Human Behaviour

Extended Data Fig. 2: Convergence Diagnostics for Random-Walk HMC Fit.

From: COVID-19 Government Response Event Dataset (CoronaNet v.1.0)

Extended Data Fig. 2: Convergence Diagnostics for Random-Walk HMC Fit.The alternative text for this image may have been generated using AI.

Plot A shows the distribution of split-Rhat values for all 40,000 parameters in the model, revealing most parameters are close to 1, which indicates strong convergence. The effective number of samples for parameters in plot B is also very high, often exceeding the total number of posterior draws. Plots C and D show strong mixing across chains for the intercept and over-time parameter for the United States for January 30th.

Back to article page