M3 Competition Dataset

该数据集名为M3，包含了3003个按类型（如工业、金融）和时间间隔（如月度、季度、年度）分类的时间序列数据。它被用作开发和比较预测方法的基准。此外，该数据集还包含了短期、中期和长期预测的时间序列数据，其结果可用于评估新预测方法的准确性。整个数据集的规模为3003个时间序列，主要任务是指向时间序列预测。This dataset, named M3, comprises 3003 time series categorized by type (e.g., industry, finance) and time interval (e.g., monthly, quarterly, annually). It serves as a benchmark for developing and comparing forecasting methods. Furthermore, the dataset includes time series data for short-, medium- and long-term forecasting, the outcomes of which can be employed to assess the accuracy of novel forecasting methods. The entire dataset consists of 3003 time series, with its primary task focusing on time series forecasting.