Statistical Analysis and Parameter Selection for Mapper
Authors: Mathieu Carrière, Bertrand Michel, Steve Oudot
JMLR 2018 | Venue PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | In this section, we provide few examples of parameter selections and confidence regions (which are unions of squares in the extended persistence diagrams) obtained with bottleneck bootstrap. ... We show in Figure 5 various Mappers ... and 85 percent confidence regions computed on various data sets. |
| Researcher Affiliation | Academia | Mathieu Carri ere EMAIL Inria Saclay 91120 Palaiseau, France Bertrand Michel EMAIL LMJL UMR 6629 Ecole Centrale Nantes 44322 Nantes, France Steve Oudot EMAIL Inria Saclay 91120 Palaiseau, France |
| Pseudocode | No | No pseudocode or algorithm block is explicitly present in the paper. The methodology is described through textual explanations and mathematical formulations. |
| Open Source Code | Yes | The code we used is available in the Gudhi open source library (see Carri ere (2017)). |
| Open Datasets | Yes | The first data set comes from the Miller-Reaven diabetes study that contains 145 observations of patients suffering or not from diabete. Observations were mapped into R5 by computing various medical features. Data can be obtained in the locfit R-package. The second data set is an instance of the 16,384-dimensional COIL data set of Nene et al. (1996). We computed the Mapper of an ant shape and a human shape from Chen et al. (2009) embedded in R3 |
| Dataset Splits | No | The paper mentions 'N = 100 subsamplings' and 'bootstrapping data 100 times' for statistical analysis and confidence region estimation, but does not provide specific train/test/validation dataset splits for model training or evaluation in the traditional machine learning sense. |
| Hardware Specification | No | The paper does not provide specific hardware details (such as GPU/CPU models, memory, or processor types) used for running the experiments. |
| Software Dependencies | No | The paper mentions using 'the Gudhi open source library (see Carri ere (2017))' but does not specify a version number for this or any other software dependency. |
| Experiment Setup | Yes | All δ parameters and resolutions were computed with Equation (9) (the δ parameters were also averaged over N = 100 subsamplings with β = 0.001), and all gains were set to 40%. |