BIAS: A Toolbox for BenchmarkingStructural Bias in the Continuous Domain -- Tests Statistics and Rejections
收藏Figshare2023-11-10 更新2026-04-08 收录
下载链接:
https://figshare.com/articles/dataset/BIAS_A_Toolbox_for_BenchmarkingStructural_Bias_in_the_Continuous_Domain_--_Tests_Statistics_and_Rejections/16546041/3
下载链接
链接失效反馈官方服务:
资源简介:
These are the raw test statistics from running the test batteries on the scenario data. - SB folder: Each subfolder corresponds to one sample size (30,50,100 and 600) either across-dimensions or per-dimension. For each parameterization of the scenarios, there is one csv-file with the test statistics.- Critical_values folder: The full set of critical values (and reference values for calculating p-values) for all settings- Rejections folder: The set of test rejections for different settings of alpha and correction method (per-dimension tests only)- RF folder: the random forests generated for predicting the type or scenario of bias based on test rejections
以下为针对场景数据运行测试套件所得的原始测试统计量。
- SB文件夹:每个子文件夹对应一种样本量(30、50、100与600),涵盖跨维度或单维度两种设置。针对场景的每种参数化方案,均对应一份包含测试统计量的CSV文件。
- Critical_values(临界值)文件夹:包含所有设置下的完整临界值集合(以及用于计算p值的参考值)。
- Rejections(拒绝结果)文件夹:包含不同显著性水平α与校正方法设置下的测试拒绝结果集合(仅针对单维度测试)。
- RF文件夹:包含基于测试拒绝结果构建、用于预测偏差类型或场景的随机森林模型。
提供机构:
van Stein, Bas; Minku, Leandro; Vermetten, Diederick; Caraffini, Fabio; Kononova, Anna V.
创建时间:
2023-11-10



