SpectroPhon DBM for ML

The dataset comprises three versions, generated for 200 subjects, 186 subjects, and 164 subjects, respectively, all derived from the extrapolated Spectrohon DBM dataset (10.17632/jt22782wjh.1). These versions include weight-based and sweat-based features for a machine learning classification tasks, with the thresholds for the four-class weight-based classification model explicitly defined in the code. Additionally, the dataset provides unseen input samples intended for use with the Gadio GUI on Hugging Face Spaces for UI testing.本数据集包含三个版本，分别面向200名、186名及164名受试者构建，所有版本均源自经外推处理的Spectrohon DBM数据集（Spectrohon DBM dataset，DOI：10.17632/jt22782wjh.1）。各版本均配备适用于机器学习分类任务的基于体重特征与基于汗液特征的特征集，其中四分类体重特征分类模型的阈值已在代码中明确定义。此外，本数据集还提供了未参与训练的输入样本，用于配合Hugging Face Spaces平台上的Gadio图形用户界面（Gadio GUI）开展UI测试工作。