Bayesian Model Selection Approach to Boundary Detection with Non-Local Priors
Authors: Fei Jiang, Guosheng Yin, Francesca Dominici
NeurIPS 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Extensive simulation studies are conducted to compare the BMS with existing methods, and our approach is illustrated with application to the magnetic resonance imaging guided radiation therapy data. |
| Researcher Affiliation | Academia | Fei Jiang Department of Statistics and Actuarial Science The University of Hong Kong feijiang@hku.hk Guosheng Yin Department of Statistics and Actuarial Science The University of Hong Kong gyin@hku.hk Dominici Francesca Harvard T.H. Chan School of Public Health Harvard University fdominic@hsph.harvard.edu |
| Pseudocode | Yes | Algorithm 1 : Screening (i) For each i in , compute Ri. (ii) If Ri , then i is selected as a candidate point. (iii) Scan through the entire data sequence, and obtain a set of Kn candidate points Hc . Algorithm 2 : Reļ¬nement Scanning (i) Compute Pr by scanning over all the candidate points in Hc (ii) For each p, obtain a set of change points corresponding to the p largest Pr 1, . . . , Kn. Optimization (iii) Select p that maximizes Pr with respect to p. |
| Open Source Code | Yes | The R code for implementing the BMS method can be downloaded from our Git Hub repository [10]. |
| Open Datasets | No | The paper describes generating synthetic data for simulations and states that the MRg RT data was provided by Dr. Zhou Shouhao, implying it is not publicly accessible. No concrete access information for a publicly available dataset is provided. |
| Dataset Splits | No | The paper does not provide specific dataset split information (e.g., exact percentages, sample counts, or citations to predefined splits) needed to reproduce the data partitioning for training, validation, or testing. |
| Hardware Specification | No | The paper does not provide specific hardware details (e.g., exact GPU/CPU models, processor types, or memory amounts) used for running its experiments. |
| Software Dependencies | No | The paper mentions "R code" but does not specify any software names with version numbers for reproducibility (e.g., specific R packages or their versions). |
| Experiment Setup | Yes | For the BMS procedure, we consider three different priors for , corresponding to the local prior, non-local moment prior and non-local inverse moment prior. Figure 2 presents the relationship between the maximum of the overand under-segmentation errors, and the value of h with sample size 1000, which indicates h 0.65 leading to the smallest segmentation error. We take the minimum distance between candidate points n I 1.5h, where h 0.5 generally works well in the simulations. |