Causal meets Submodular: Subset Selection with Directed Information
Authors: Yuxun Zhou, Costas J. Spanos
NeurIPS 2016 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We evaluate the theoretical results with several case studies, and also illustrate the application of the subset selection to causal structure learning. ... Finally in Section 5, we conduct experiments to justify our theoretical findings and illustrate a causal structure learning application. |
| Researcher Affiliation | Academia | Yuxun Zhou Department of EECS UC Berekely yxzhou@berkeley.edu Costas J. Spanos Department of EECS UC Berkeley spanos@berkeley.edu |
| Pseudocode | Yes | Algorithm 1 Random Greedy for Subset Selection |
| Open Source Code | No | The paper does not contain any explicit statements about releasing source code for the described methodology, nor does it provide a link to a code repository. |
| Open Datasets | No | The synthesis data is generated with the Bayes network Toolbox (BNT) [15] using dynamic Bayesian network models. Two sets of data, denoted by D1 and D2, are simulated, each containing 15 and 35 processes, respectively. ... The stock market dataset, denoted by ST, contains hourly values of 41 stocks and indexes for the years 2014-2015. (The paper describes the datasets used, including synthetic data generated with a toolbox and a stock market dataset, but it does not provide concrete access information such as a link, DOI, or a citation to a publicly available version of these specific datasets.) |
| Dataset Splits | No | The paper does not provide specific details on training, validation, or test dataset splits, such as percentages, sample counts, or citations to predefined splits. |
| Hardware Specification | No | The paper does not provide specific hardware details (such as GPU or CPU models, processor types, or memory amounts) used for running its experiments. |
| Software Dependencies | No | The paper mentions the use of 'Bayes network Toolbox (BNT) [15]' but does not provide specific version numbers for BNT or any other software dependencies, such as programming languages or libraries, which would be necessary for reproduction. |
| Experiment Setup | Yes | The maximal context tree depth is set to 5, which is sufficient for both the synthesis datasets and the real-world ST dataset. ... The order (memory length) of the historical dependence is set to 3. The MCMC sampling engine is used to draw n = 104 points for both D1 and D2. ... we detrend each time series with a recursive HP-filter [24] to remove long-term daily or monthly seasonalities that are not relevant for hourly analysis. |