Title: The Quality of the Covariance Selection Through Detection Problem and AUC Bounds

Abstract: We consider the problem of quantifying the quality of a model selection problem for a graphical model. We discuss this by formulating the problem as a detection problem. Model selection problems usually minimize the distance with the model distribution. For the special case of Gaussian distributions, this problem simplifies to the covariance selection problem which is widely discussed in literature by Dempster [1] where the Kullback-Leibler (KL) divergence is minimized or equivalently the likelihood criterion maximized to compute the model covariance matrix. While this solution is optimal for Gaussian distributions in the sense of the KL divergence, it is not optimal when compared with other information divergences and criteria such as Area Under the Curve (AUC).
In this paper, we discuss the quality of model approximation using the AUC and its bounds as an average measure of accuracy in detection problem. We compute upper and lower bounds for the AUC. We define the correlation approximation matrix (CAM) and show that the KL divergence and AUC and its upper and lower bounds depend on the eigenvalues of the CAM. We also show the relationship between the AUC, the KL divergence and the ROC curve by optimizing with respect to the ROC curve. In the examples provided, we pick tree structures as the simplest graphical models. We perform simulations on fully-connected graphs and compute the tree structured models by applying the widely used Chow-Liu algorithm. Examples show that the quality of tree approximation models are not good in general based on information divergences, AUC and its bounds when the number of nodes in the graphical model is large. Specially for 1-AUC, it is shown both in theory and using simulations that the 1-AUC for the tree approximation model decays exponentially as the dimension of the graphical model increases.