GemBag: Group Estimation of Multiple Bayesian Graphical Models

Authors: Xinming Yang, Lingrui Gan, Naveen N. Narisetty, Feng Liang

JMLR 2021 | Venue PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Through extensive simulation studies and an application to a bike sharing data set, we demonstrate that the proposed Gem Bag procedure has strong empirical performance in comparison with alternative methods.
Researcher Affiliation Collaboration Xinming Yang EMAIL Department of Statistics University of Illinois at Urbana-Champaign Champaign, IL, USA Lingrui Gan EMAIL Facebook Menlo Park, CA, USA Naveen N. Narisetty EMAIL Department of Statistics University of Illinois at Urbana-Champaign Champaign, IL, USA Feng Liang EMAIL Department of Statistics University of Illinois at Urbana-Champaign Champaign, IL, USA
Pseudocode Yes Algorithm 1 EM algorithm for computing the MAP estimator in (12)
Open Source Code No The paper does not provide concrete access to source code for the methodology described. It provides a link to the journal page for the paper and its license, but no specific code repository or explicit code release statement.
Open Datasets Yes We use the Capital Bikeshare trip data1 to evaluate the performance of Gem Bag. The data contains records of bike rentals by either a registered rider or a casual rider in a bicycle sharing system with more than 500 stations. ... 1. Data available at https://www.capitalbikeshare.com/system-data.
Dataset Splits Yes For each class, we use the first 80 percent observations as training data and the rest 20 percent as test data.
Hardware Specification Yes We report the computational time of Gem Bag along with the competing methods using a Mac Book Pro with 2.9 GHz Intel Core i5 processor and 8.00 GB memory in Table 5.
Software Dependencies No The paper mentions 'Map built using the R package ggmap (Kahle and Wickham, 2013)' but does not provide specific version numbers for R or the ggmap package, nor for any other key software components used for the main models or experiments.
Experiment Setup Yes For Gem Bag, we consider α 1, ?n, and n and estimate the sparsity patterns using a threshold t 0.5 on the posterior inclusion probabilities. We set p1 p2 ? 0.5 so that the prior inclusion probability Pprk,ij 1q p1p2 0.5 and tune pv0, v1q with v0 τ p0.25, 0.5, 0.75, 1q ˆ a 1{pn log pq and v1 p2.5, 5, 7.5, 10q ˆ a 1{pn log pq when α 1, with v0 τ p1, 1.5, 2, 2.5, 3qˆ10 2ˆ a 1{log p and v1 p2, 4, 6, 8qˆ a 1{log p when α ?n, or with v0 τ p1, 1.5, 2, 2.5, 3qˆ10 3ˆ a n{log p and v1 p2, 4, 6, 8qˆ a n{log p when α n.