High-Dimensional Structured Quantile Regression
Authors: Vidyashankar Sivakumar, Arindam Banerjee
ICML 2017 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We perform experiments on synthetic data which support the theoretical results. |
| Researcher Affiliation | Academia | Department of Computer Science & Engineering, University of Minnesota, Twin Cities. |
| Pseudocode | No | The paper does not contain any pseudocode or algorithm blocks. |
| Open Source Code | No | The paper does not provide any concrete access to source code for the methodology described. |
| Open Datasets | No | Data is generated as y = Xθ + ω... We perform simulations with synthetic data. |
| Dataset Splits | No | The paper does not provide specific dataset split information (exact percentages, sample counts, or citations to predefined splits) needed to reproduce the data partitioning. It describes the generation of synthetic data and varying sample sizes (n) for simulations. |
| Hardware Specification | No | The paper does not provide specific hardware details (e.g., GPU/CPU models, memory amounts) used for running its experiments. It only states "The code was implemented in Python." |
| Software Dependencies | No | The paper states: "For the optimization, we use the Alternating Direction Method of Multipliers (Boyd et al., 2010). The details of the updates can be found in the flare documentation Li et al. (2015). The code was implemented in Python." However, it does not provide specific version numbers for Python or the 'flare' package. |
| Experiment Setup | Yes | Data is generated as y = Xθ + ω. θ = [1, 1, 1, 1, 1, 1 | {z } 6 , 0, 0, . . . , 0 | {z } p-6 ] Rp for the l1 norm and θ = [1, . . . , 1 | {z } 5 , 1, . . . , 1 | {z } 5 , 1, . . . , 1 | {z } 5 , 0, . . . , 0 | {z } 5 , . . . , 0, . . . , 0 | {z } 5 for the l1/l2 group sparse norm with p [500, 750, 1000]. The noise ωi N(0, 0.25), i [n] is Gaussian with zero mean and 0.25 variance. The design matrix X N(0, Ip p) is multivariate Gaussian with identity covariance. We vary n = [10, 20, 30, . . . , 120, 130]. For each n we generate 100 datasets with the probability of success defined as the fraction of times we are able to faithfully estimate the true parameter. For p = 500 we run simulations for τ [0.1, 0.5, 0.9] and for p [750, 1000] we run simulations only for τ = 0.5. For the optimization, we use the Alternating Direction Method of Multipliers (Boyd et al., 2010). |