The content of this page refers to an obsolete version of BayesiaLab. Please see the up-to-date entry for Structural Coefficient Analysis (opens in a new tab).
If both Learning and Test Sets are available, a Validation Measure can be computed to help chose the most appropriate Structural Coefficient.
This measure is based on the Test Set’s mean negative log-likelihood (returned by the network learned from the Learning Set) and on the variances of the negative log-likelihood of the Test Set and Learning Set (returned by the network learned from Learning Set).
The "Bathtub"
The bottom portion of the "bathtub" suggests a range of Structural Coefficients for models that do neither overfit nor underfit the data.
