Figure 6

Comparing the Jensen Shannon divergence to the Bayes factor updates for regression models. Top row: Solid gray line indicate the ’true’ system response to the true parameters which is the same in each graph. The solid black line indicates the data generated by model three, which was used to determine the posterior predictive distribution on the four models. Dashed lines indicate the inferred credible intervals of the Posterior Predictive Distribution for the model corresponding to that column. Second row: Mean of the predicted change in Bayes factor in favor of model three after incorporating two additional data points from the posterior predictive distribution of the true model (average of 100 repetitions). Coordinate along each axis reflects the point t at which the measurement is performed. Bottom row: Jensen Shannon Divergence between the Posterior Predictive Distributions of M ₃ and the model that column number corresponds to. Here M ₃ corresponds to the model which generated the data. Note that the graphs for model three are black by definition.