Variable importance excluding outcome variables as predictors.

Figshare2025-02-06 更新2026-04-28 收录

数据链接：

https://figshare.com/articles/dataset/Variable_importance_excluding_outcome_variables_as_predictors_/28363210 数据链接链接失效反馈

官方服务：

资源简介：

Variable importance excluding outcome variables as predictors.

创建时间：

2025-02-06

Straight line, 4D AND, 5D XOR, Hypersphere, Cone, Trigonometric, Double Spiral, Yin-yang, 5 multi-cut, 10 multi-cut

特征选择

数据集成

用于基准测试特征选择算法的合成数据集，包括多种类型的数据集，如直线、4D AND、5D XOR等，每个数据集都有其特定的特征和目标变量方程。

github2024-05-23 更新150

SolshineMisfit/nla-qwen2.5-7b-L20-results-famous-ads

模型可解释性

广告分析

该数据集名为著名广告标语——NLA激活分析（Qwen2.5-7B，第20层），旨在探究语言模型在读取著名广告标语时的内部表示。它使用自然语言自编码器（NLA）将Qwen2.5-7B-Instruct模型在第20层最后一个token的残差流激活解码为自然语言描述。数据集包含约146条著名广告标语（源自维基百科，附有品牌、年份和上下文信息），每条标语都对应一个由AV生成的英文解释（描述模型对该标语的表

Hugging Face2026-05-27 更新70

Regression coefficient was obtained by L1-penalized logistic regression (details are described in Material and Methods) and sorted by descending absolute values.

L1正则化

特征选择

Elastic-net parameters alpha was set to 0.975 and lambda was estimated as 0.075. Coefficients were calculated with standardized variables (STDV = 1).

NIAID Data Ecosystem40

The highest classification accuracy of each algorithm on each data set and its feature subset length.

机器学习

特征选择

The highest classification accuracy of each algorithm on each data set and its feature subset length.

NIAID Data Ecosystem80

Importance coefficient and ranking of features in LDA model.