EmotiBit生理信号数据集
收藏资源简介:
该数据集由28名参与者完成24个陈述评估任务,共672个试验组成。每个试验都记录了参与者的皮肤电导(EDA)和周围血流(PPG)信号,并标注了参与者是否相信陈述以及他们是否之前见过这个陈述。该数据集为研究对虚假信息的生理反应提供了一个新的资源,并使用机器学习模型分析了这些生理模式。结果表明,皮肤电导(EDA)在预测准确率方面优于PPG。这些发现表明,可穿戴传感器作为一种最小的侵入性工具,在检测信念和先前接触方面具有潜力,为实时虚假信息检测和自适应、用户感知系统提供了新的方向。
This dataset comprises 672 trials generated from 24 statement evaluation tasks completed by 28 participants. For each trial, the participants' skin conductance (EDA) and peripheral blood flow (PPG) signals were recorded, alongside annotations indicating whether the participant believed the given statement and whether they had previously encountered the statement. This dataset provides a novel resource for researching physiological responses to misinformation, and machine learning models were employed to analyze these physiological patterns. The results demonstrate that skin conductance (EDA) outperforms PPG in terms of prediction accuracy. These findings suggest that wearable sensors, as minimally invasive tools, hold potential for detecting belief and prior exposure, opening new directions for real-time misinformation detection and adaptive, user-aware systems.




