Graphical model structure change

A model for detecting graphical model structure changes was applied to some simulated data for which we know the real changes.

Usage

data(change)

Format

List of 4 data.frames: coefs contains the coefficients for every variable at every point in the regularization path, truth contains information about which variables are real changes, likelihood contains the train and test likelihood, and position contains positions for laying out the graphical model nodes on the 2D plane.

Source

Song Liu, personal communication.

References

Liu S, Quinn JA, Gutmann MU, Sugiyama M. Direct Learning of Sparse Changes in Markov Networks by Density Ratio Estimation. ECML/PKDD 2013.