Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in [1].
Markov Blanket and Markov Boundary of Multiple Variables
Authors: Xu-Qing Liu, Xin-Sheng Liu
JMLR 2018 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Finally, we make a benchmarking study based on six synthetic BNs and then apply MB discovery to multi-class prediction based on a real data set. The experimental results reveal our algorithms have higher accuracies and lower complexities than existing algorithms. |
| Researcher Affiliation | Academia | Xu-Qing Liu EMAIL State Key Laboratory of Mechanics and Control of Mechanical Structures Institute of Nano Science and Department of Mathematics Nanjing University of Aeronautics and Astronautics, Nanjing 210016, China Faculty of Mathematics and Physics Huaiyin Institute of Technology, Huai an 223003, China Xin-Sheng Liu EMAIL State Key Laboratory of Mechanics and Control of Mechanical Structures Institute of Nano Science and Department of Mathematics Nanjing University of Aeronautics and Astronautics, Nanjing 210016, China |
| Pseudocode | Yes | Algorithm 1: IAMBS and KIAMBS |
| Open Source Code | No | The paper does not provide explicit statements or links for the open-sourcing of the authors' developed algorithms (MIAMB, MKIAMB, IAMBS, KIAMBS). It mentions using third-party toolboxes like Full BNT (Murphy, 2007) and MIToolbox (Brown et al., 2012) and provides a link for Lib SVM (Chang and Lin, 2011) which is a classifier they used. |
| Open Datasets | Yes | We use the data sets of sizes 500 and 5000, generated by Tsamardinos et al. (2006) and Aliferis et al. (2010a), which are available at http://www.nyuinformatics.org/ downloads/supplements/JMLR2009/index.html. Data: HIVA contains 4229 data points and 1618 variables. This data set is very challenging in WCCI 2006 (http://www.modelselect.inf.ethz.ch) and IJCNN 2007 (http://www.agnostic.inf.ethz.ch) |
| Dataset Splits | Yes | All the classifications are performed by 10-fold cross-validation. |
| Hardware Specification | Yes | Here, RT refers to the single CPU time implemented on an Intel i7-3612QM 2.1 GHz and Windows 7 with 64 bits. |
| Software Dependencies | Yes | According to the above description, we make computations with the aid of Full BNT (Murphy, 2007) and MIToolbox (Brown et al., 2012). ... the support vector machines (SVMs; implemented via Lib SVM v3.22) |
| Experiment Setup | Yes | We take K = 0.8 as the randomization parameter in KIAMB and MKIAMB due to the following two reasons: (i) Pe na et al. (2007, p. 227) asserted that K [0.7, 0.9] performs best; and (ii) K = 0.8 is an appropriate tradeoffbetween WA (or WP) and RT. ... we used the damped G2 test by setting κ = 5. |