FIGURE 8 |.

Stability test of the displacement (trajectories) of the model prediction (in blue) and the model average (in red) under deletion of 1 – 30 models. M denotes the true location of the orthogonal projection of the generating model in the hyperplane. m and a mark the location of the model projection and the model average, respectively, when the 30 models are used. In both cases, as models are removed one by one from the candidate model set, the location of both m and a changes (little vertical lines). Note how the model projection estimate is more stable to changes in the model set than the model average. Prasanta S. Bandyopadhyay, Gordon Brittan Jr., Mark L. Taper, Belief, Evidence, and Uncertainty. Problems of Epistemic Inference, published 2016 Springer International Publisher, reproduced with permission of Springer Nature Customer Service Center.