A0215
Title: Network detection through odds ratio model
Authors: Jinsong Chen - University of Nevada Reno (United States) [presenting]
Abstract: A unified modelling approach to network detection through the semi-parametric odds ratio model is proposed. The proposed model is flexible in handling discrete and continuous data, invariant to biased sampling designs, and avoids model incompatibility. A neighbourhood selection approach is proposed and is shown to be sign-consistent under a version of the irrepresentable condition. A coordinate descent algorithm is proposed to solve the computation problem. Simulations demonstrate that the proposed approach has good performance in comparison to the Gaussian modelling approach. The proposed approach is applied to detect the gene-expression network in breast invasive carcinoma.