A*HAR

A*HAR数据集是由新加坡科技研究局信息通信研究所创建，旨在解决类别不平衡的人类活动识别问题。该数据集包含24种日常活动，如散步、跑步、举重等，数据量达24,770条，使用手腕穿戴传感器收集。创建过程中，通过MetaBase应用进行数据收集和标注。A*HAR数据集特别适用于研究半监督学习算法，尤其是在处理类别不平衡问题上的应用。The A*HAR Dataset was developed by the Institute for Infocomm Research under the Agency for Science, Technology and Research (A*STAR) of Singapore, aiming to address the class imbalance problem in human activity recognition. This dataset includes 24 daily activities such as walking, running, weightlifting and others, with a total of 24,770 data samples collected using wrist-worn sensors. During the dataset creation process, data collection and annotation were carried out via the MetaBase application. The A*HAR Dataset is particularly suitable for researching semi-supervised learning algorithms, especially their applications in handling class imbalance issues.