Direct Estimation of Differential Functional Graphical Models

Authors: Boxin Zhao, Y. Samuel Wang, Mladen Kolar

NeurIPS 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We illustrate finite sample properties of our method through simulation studies. Finally, we apply our method to EEG data to uncover differences in functional brain connectivity between alcoholics and control subjects.
Researcher Affiliation Academia Boxin Zhao Department of Statistics The Unveristy of Chicago Chicago, IL 60637 boxinz@uchicago.edu Y. Samuel Wang Booth School of Business The Unveristy of Chicago Chicago, IL 60637 swang24@uchicago.edu Mladen Kolar Booth School of Business The Unveristy of Chicago Chicago, IL 60637 mkolar@chicagobooth.edu
Pseudocode Yes Algorithm 1 Functional differential graph estimation
Open Source Code Yes 1The code for this part is on https://github.com/boxinz17/Fu DGE
Open Datasets Yes We apply our method to electroencephalogram (EEG) data obtained from an alcoholism study [29, 6, 18]
Dataset Splits Yes Both M and L are chosen by 5-fold cross-validation as discussed in [18].
Hardware Specification No No specific hardware details (e.g., GPU, CPU models, memory) were mentioned for running experiments.
Software Dependencies No The paper mentions functional data analysis techniques (e.g., FPCA, B-spline basis) and optimization methods (e.g., proximal gradient method, group lasso), but does not specify any software libraries or packages with version numbers used for implementation.
Experiment Setup Yes In each setting, we generate n X p functional variables from graph GX via Xij(t) = b(t)T δX ij , where b(t) is a five dimensional basis with disjoint support over [0, 1] with bk(t) = cos (10π (x (2k 1)/10)) + 1 (k 1)/5 x < k/5; 0 otherwise, k = 1, . . . , 5. ... We choose λn so that the estimated differential graph has approximately 1% of possible edges. The estimated edges of the differential graph are shown in Figure 3.