Uncertainty Quantification for Data-Driven Change-Point Learning via Cross-Validation
Authors: Hui Chen, Yinxu Jia, Guanghui Wang, Changliang Zou
AAAI 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | In this section, we demonstrate the performance of the proposed methods through numerical studies. The experiments are run on a personal computer with an Intel Core i7-10700 CPU, 8GB of memory, a 64-bit operating system and R software version 4.2.1. Table 2 presents a comparative analysis between our proposed method and the benchmark approach SMUCE under normal and t(5) error term distributions. Our method consistently maintains P+ within the specified α for both distributions. |
| Researcher Affiliation | Academia | 1School of Mathematics and Statistics, Jiangsu Normal University 2School of Statistics and Data Science, LPMC, KLMDASR, and LEBPS, Nankai University 3School of Statistics, Academy of Statistics and Interdisciplinary Sciences, and KLATASDS-MOE, East China Normal University |
| Pseudocode | No | The paper describes algorithmic steps within the 'Testing Procedure' and 'Refinement with Cross-Validation' sections, but it does not include a clearly labeled pseudocode block or algorithm figure. |
| Open Source Code | No | The paper does not provide any statement or link indicating that the source code for the proposed methodology is openly available. |
| Open Datasets | Yes | Consider the array comparative genomic hybridization (CGH) data from the coriell dataset available in the R package DNAcopy in Seshan and Olshen (2019). |
| Dataset Splits | Yes | Inspired by the order-preserved splitting strategy proposed by Zou, Wang, and Li (2020), we partition the data into training set ξtr and validation set ξte with ntr and nte observations respectively, according to whether the observed index being odd or even, such that the two data sets share a similar change-point pattern as much as possible. We also investigate the refinement of our method using multiple folds cross-validation (RC) as discussed in the methodology section. For illustration purposes, we conduct the experiments using a three-fold cross-validation approach. |
| Hardware Specification | Yes | The experiments are run on a personal computer with an Intel Core i7-10700 CPU, 8GB of memory, a 64-bit operating system and R software version 4.2.1. |
| Software Dependencies | Yes | The experiments are run on a personal computer with an Intel Core i7-10700 CPU, 8GB of memory, a 64-bit operating system and R software version 4.2.1. We use the implementations of these algorithms provided in the R packages wbs (Baranowski and Fryzlewicz 2019) and changepoint (Killick, Haynes, and Eckley 2016; Killick and Eckley 2014). |
| Experiment Setup | Yes | All the simulation results are based on 500 replications and the bootstrap sample size is B = 500. The nominal level is α = 10%. (from Table 1 caption) SNR= 1.2 (from Table 2 caption). |