LEAD1.0
收藏资源简介:
LEAD1.0是由新加坡管理大学和Berkeley Education Alliance for Research in Singapore Limited共同创建的大型能源异常检测数据集,包含1,413个智能电表的一年时间序列数据。该数据集通过手动标注,区分了点异常和序列异常,旨在帮助研究者开发和评估建筑物能源消耗中的异常检测技术。数据集的创建过程涉及对约1200万数据点的仔细检查和标注,应用了基于时间窗口的固定协议。该数据集的应用领域主要集中在通过数据驱动的分析技术减少建筑物能源浪费,提高能源使用效率,从而推动全球能源可持续发展。
LEAD1.0 is a large-scale energy anomaly detection dataset jointly developed by Singapore Management University and Berkeley Education Alliance for Research in Singapore Limited. It encompasses one-year time-series data collected from 1,413 smart meters. This dataset is manually annotated to distinguish between point anomalies and sequential anomalies, with the core objective of assisting researchers in developing and evaluating anomaly detection technologies for building energy consumption. The creation of this dataset involved meticulous inspection and annotation of approximately 12 million data points, adhering to a fixed time-window-based protocol. Its main application areas focus on reducing building energy waste and improving energy utilization efficiency via data-driven analytical technologies, thereby advancing the sustainable development of global energy.




