Causal Discovery from Nonstationary/Heterogeneous Data: Skeleton Estimation and Orientation Determination
Authors: Kun Zhang, Biwei Huang, Jiji Zhang, Clark Glymour, Bernhard Schölkopf
IJCAI 2017 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Experimental results on various synthetic and real-world data sets are presented to demonstrate the efficacy of our methods. |
| Researcher Affiliation | Academia | Department of philosophy, Carnegie Mellon University MPI for Intelligent Systems, T ubingen, Germany Lingnan University, Hong Kong {kunz1,biweih}@andrew.cmu.edu, jijizhang@ln.edu.hk, cg09@andrew.cmu.edu, bs@tuebingen.mpg.de |
| Pseudocode | Yes | Algorithm 1 Detection of Changing Modules and Recovery of Causal Skeleton |
| Open Source Code | No | The paper does not provide an explicit statement about the release of source code for the described methodology or a direct link to a code repository. |
| Open Datasets | Yes | This f MRI Hippocampus dataset [Poldrack and Laumann, 2015] contains signals from six separate brain regions... The breast tumor dataset is from the UCI Machine Learning Depository [Blake and Merz, 1998]. |
| Dataset Splits | Yes | we trained SVM with these 11 features, subsets of these 11 features, random subsets of all features, and all 30 features, and used 10-fold cross-validation (CV) error to assess the classification accuracy. |
| Hardware Specification | No | The paper does not provide specific details on the hardware used for running experiments, such as CPU/GPU models or memory specifications. |
| Software Dependencies | No | The paper mentions using specific methods like 'kernel-based conditional independence test (KCI-test [Zhang et al., 2011])' and 'SGS search [Spirtes et al., 1993]', but does not provide specific version numbers for any software dependencies or libraries used in the implementation. |
| Experiment Setup | Yes | We tried different periods (w = 5, 10, 20, 30) on the time-varying component a, as well as different sample sizes (N = 600, 1000). In each setting, we ran 50 trials. ... We applied our enhanced constraint-based method on 10 successive days separately, with time information T as an additional variable in the system. ... we trained SVM with these 11 features, subsets of these 11 features, random subsets of all features, and all 30 features, and used 10-fold cross-validation (CV) error to assess the classification accuracy. |