Skip to main content
Fig. 3 | The Journal of Mathematical Neuroscience

Fig. 3

From: Measuring Edge Importance: A Quantitative Analysis of the Stochastic Shielding Approximation for Random Processes on Graphs

Fig. 3

Mean squared errors for the 3-state chain. Comparison of mean squared errors of M ⊺ U ( i , j ) in the 3-state chain, i.e. the projection of U ( i , j ) onto the third component. The theoretical MSE values are computed by summing the appropriate edge importance values R k . M ⊺ U ( 1 , 2 ) has the smallest MSE out of the three classes of 2-noise source reduced process, showing that, as observed by Schmandt and Galán, X ( 1 , 2 ) is optimal at preserving the accuracy of the full process with respect to the third component of the system

Back to article page