SWELL-KW

该网站介绍了新的多模式膨胀知识工作 (SWELL-KW) 数据集，用于研究压力和用户建模。该数据集是在一个实验中收集的，其中25个人进行了典型的知识工作 (撰写报告，进行演示，阅读电子邮件，搜索信息)。我们用压力来操纵他们的工作条件: 电子邮件中断和时间压力。记录了各种数据: 计算机日志记录，摄像机记录的面部表情，Kinect 3D传感器的身体姿势以及人体传感器的心率 (可变性) 和皮肤电导。我们的数据集不仅包含原始数据，还包含预处理数据和提取的特征。通过经过验证的问卷作为基本事实，评估了参与者在任务负荷，精神努力，情绪和感知压力方面的主观经验。生成的有关工作行为和情感的数据集适用于多个研究领域，例如工作心理学，用户建模和上下文感知系统。

This website presents a novel multimodal SWELL-KW dataset for stress and user modeling research. This dataset was collected via an experiment where 25 participants completed typical knowledge work tasks, including report writing, presentation preparation, email reading, and information searching. We manipulated their work conditions to induce stress through two factors: email interruptions and time pressure. Various types of data were recorded: computer logs, facial expressions captured by video cameras, body postures collected via Kinect 3D sensors, as well as heart rate (and heart rate variability) and skin conductance measured by body-worn physiological sensors. In addition to raw data, the dataset also includes preprocessed data and extracted features. Participants' subjective experiences regarding task load, mental effort, emotional state, and perceived stress were evaluated using validated questionnaires as ground truth. This dataset focusing on work behaviors and emotions is applicable to multiple research domains, such as occupational psychology, user modeling, and context-aware systems.

搜集汇总

背景与挑战

背景概述

SWELL-KW是一个多模式数据集，专门用于研究知识工作中的压力和用户建模，通过实验收集了25名参与者在受控工作条件下的计算机日志、生理信号和视频数据，并包含预处理特征和主观问卷评估。该数据集适用于工作心理学、用户建模和上下文感知系统等多个研究领域，由代尔夫特理工大学和Radboud University于2014年发布。