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; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in [1]
Inference on the Change Point under a High Dimensional Covariance Shift
Authors: Abhishek Kaul, Hongjin Zhang, Konstantinos Tsampourakis, George Michailidis
JMLR 2023 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Easy to implement algorithms for the proposed methodology are developed and their performance illustrated on synthetic and real data sets. ... Evaluation of Estimation and Inference Properties: Next, an evaluation of the proposed estimation and inference methodology is provided on synthetic data. ... All simulation experiments in Section 4 consider a preliminary search grid... |
| Researcher Affiliation | Academia | Abhishek Kaul EMAIL Department of Mathematics and Statistics Washington State University Pullman, WA 99164, USA. Hongjin Zhang EMAIL Department of Mathematics and Statistics Washington State University Pullman, WA 99164, USA. Konstantinos Tsampourakis EMAIL School of Mathematics University of Edinburgh Edinburgh, Scotland, EH9 3FD. George Michailidis EMAIL Department of Statistics and the Informatics Institute University of Florida Gainsville, FL 32611-8545. |
| Pseudocode | Yes | Algorithm 1: Op(ψ 2) estimation of τ 0 : ... Algorithm 2: Op(ψ 2) estimation of τ 0 : ... Algorithm 3: Empirically fitting a centered and scaled χ2 k to the distribution law L. ... Algorithm 4: Overall Inference Procedure for τ 0 |
| Open Source Code | No | The paper mentions that "All computations are carried out in R, and all Lasso optimizations of 3.1 are carried out using the glmnet package." and references the "R-package Changeponts HD" for a benchmark method. However, there is no explicit statement or link indicating that the authors' own code for the proposed methodology is open-source or publicly available. |
| Open Datasets | Yes | The Global Gut data set contains measurements of individuals from several geographical locations...For our analyses we consider the publicly available global human gut microbiome data of Yatsunenko et al. (2012). |
| Dataset Splits | No | The paper discusses simulation parameters such as "The observation period T is set to {300, 400, 500}, the dimension p to {25, 50, 150, 250} and the relative location of the change point (τ 0/T) {0.2, 0.4, 0.6, 0.8}" and a "preliminary search grid of ˇτ {0.25, 0.5, 0.75}". For the real data, it mentions processing steps like "subset the analysed set of genera" and "log-relative abundance transformation". However, it does not provide specific train/test/validation splits for any dataset used in the experiments. |
| Hardware Specification | No | The paper mentions that "All computations are carried out in R" and uses the "glmnet package" and "R-package Changeponts HD". However, it does not provide any specific details about the hardware used to run the experiments, such as CPU or GPU models, or memory specifications. |
| Software Dependencies | No | The paper states: "All computations are carried out in R, and all Lasso optimizations of 3.1 are carried out using the glmnet package." It also mentions comparing to "the method of Bybee and Atchad e (2018)" using "its implementation... via the authors developed R-package Changeponts HD". While it names software, it does not provide specific version numbers for R, glmnet, or Changeponts HD. |
| Experiment Setup | Yes | The observation period T is set to {300, 400, 500}, the dimension p to {25, 50, 150, 250} and the relative location of the change point (τ 0/T) {0.2, 0.4, 0.6, 0.8}. ... The significance level is set to α {0.05, 0.01} in all cases. ... To construct Σ, we consider a Toeplitz type matrix Γ with the (l, m)th component set as Γ(l,m) = ρ|l m|a, l, m = 1, ..., p. We set ρ = 0.4 and a = 1/ log s, where s specified below. ... The tuning parameters λj, j = 1, ..., p used to obtain ℓ1 regularized mean estimates are selected based on a BIC type criterion. Specifically, we set λj = λ, j = 1, ..., p, and evaluate ˆµ(j)(λ), and ˆγ(j)(λ) over an equally spaced grid of seventy five values in the interval (0, 1). |