Post-Regularization Inference for Time-Varying Nonparanormal Graphical Models
Authors: Junwei Lu, Mladen Kolar, Han Liu
JMLR 2017 | Venue PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Our method is supported by thorough numerical simulations and an application to a neural imaging data set. ... We illustrate finite sample properties of the estimator in (3) on synthetic data. ... We apply the super graph test in Section 3.2 to the ADHD-200 brain imaging data set (Biswal et al., 2010). |
| Researcher Affiliation | Academia | Junwei Lu EMAIL Department of Operations Research and Financial Engineering Princeton University Princeton, NJ 08544, USA. Mladen Kolar EMAIL Booth School of Business The University of Chicago Chicago, IL 60637, USA. Han Liu EMAIL Department of Operations Research and Financial Engineering Princeton University Princeton, NJ 08544, USA |
| Pseudocode | No | The paper describes the inferential methods and estimation procedures using mathematical formulas and textual descriptions, but it does not include any clearly labeled 'Pseudocode' or 'Algorithm' blocks. |
| Open Source Code | No | The paper includes a license statement and attribution requirements for the journal article itself ('License: CC-BY 4.0, see https://creativecommons.org/licenses/by/4.0/. Attribution requirements are provided at http://jmlr.org/papers/v18/17-145.html.'), but it does not contain any explicit statement about releasing the source code for the methodology described in the paper, nor does it provide a link to a code repository. |
| Open Datasets | Yes | Our method is supported by thorough numerical simulations and an application to a neural imaging data set. ... We apply the super graph test in Section 3.2 to the ADHD-200 brain imaging data set (Biswal et al., 2010). The ADHD-200 data set is a collection of resting-state functional MRI (R-f MRI) of subjects with and without attention deficit hyperactive disorder (ADHD) (Eloyan et al., 2012; The ADHD-200 Consortium, 2012). ... We focus on 264 voxels as the regions of interest (ROI) extracted by Power et al. (2011). |
| Dataset Splits | No | For the synthetic data, the paper describes the data generation process and samples used ('sample size n {200, 500, 800}') and measures overall TPR and FPR, but it does not specify explicit training, validation, or test splits for model evaluation. For the ADHD-200 data, it describes the data as '776 subjects: 491 controls, 195 cases diagnosed with ADHD ... and 88 subjects with withheld diagnosis' and focuses on '491 controls' for analysis, but it does not detail any splits for model training or testing. |
| Hardware Specification | No | The Acknowledgments section mentions general computing resources: 'This work was completed in part with resources provided by the University of Chicago Research Computing Center.' However, it does not specify any particular CPU, GPU, or other hardware models used for the experiments. |
| Software Dependencies | No | The paper describes various estimators like 'kernel Pearson CLIME estimator', 'kernel graphical Lasso estimator', and 'kernel neighborhood selection estimator', but it does not specify any software libraries, packages, or solvers with version numbers that were used for their implementation or experiments. |
| Experiment Setup | Yes | The bandwidth parameter is set as h = 0.35/n1/5 for all four methods in order to facilitate easier comparison. For the calibrated CLIME, we choose γ = 0.5 in (3). ... we set the tuning parameters for our procedure as γ = 0.5, the bandwidth h = 0.002 and the penalty parameter λ = 0.03 h2 + p d/h/(nh), where the last two parameters are chosen through cross-validation. |