hugging-science/AI4FA-Goldberg

这是一个实验性的微生物组数据集，设计用于食物过敏分类的探索性研究。数据集包含多种数据模态（DNA嵌入、微生物组嵌入、原始DNA序列），这些数据是在多个时间点（1、2、3、6、12、24、36个月）纵向收集的。数据集的结构按照数据类型和时间点组织，包括DNA嵌入（.h5文件）、微生物组嵌入（.h5文件）和DNA序列（.csv文件）。每个时间点包含每个受试者的多个样本。数据集的用途是探索性的食物过敏分类，用户需要自行定义训练/测试分割，并注意在受试者级别进行分割以避免数据泄露。数据集是纵向和多模态的，但没有嵌入标签，需要用户自行处理或映射。

This is an experimental microbiome dataset designed for exploratory research in food allergy classification. The dataset contains multiple data modalities (DNA embeddings, microbiome embeddings, raw DNA sequences) collected longitudinally at several timepoints (1, 2, 3, 6, 12, 24, 36 months). The dataset is organized by data type and timepoint, including DNA embeddings (.h5 files), microbiome embeddings (.h5 files), and DNA sequences (.csv files). Each timepoint contains multiple samples per subject. The intended use is exploratory classification of food allergies, and users are expected to define their own train/test splits, ensuring splits are done at the subject level to avoid data leakage. The dataset is longitudinal and multi-modal, but no labels are embedded per file, requiring users to handle or map them separately.