ETT (Electricity Transformer Temperature)

电力变压器温度 (ETT) 是电力长期部署的关键指标。该数据集由来自中国两个分离县的2年数据组成。为了探索长序列时间序列预测 (LSTF) 问题的粒度，创建了不同的子集，{ETTh1，ETTh2} 为1小时级，ETTm1为15分钟级。每个数据点由目标值 “油温” 和6个功率负载特征组成。火车/val/测试为12/4/4个月。

Electrical Transformer Temperature (ETT) is a critical metric for the long-term deployment of power systems. This dataset consists of 2 years of data collected from two separate counties in China. To explore the granularity of the Long Sequence Time-series Forecasting (LSTF) problem, multiple subsets are created: ETTh1 and ETTh2 are at 1-hour granularity, while ETTm1 is at 15-minute granularity. Each data point comprises the target variable "oil temperature" and six power load features. The training, validation, and test sets correspond to 12, 4, and 4 months of data respectively.

AI搜集汇总

背景与挑战

背景概述

ETT数据集是一个用于长序列时间序列预测的电力变压器温度数据集，包含来自中国两个县的两年数据，目标变量为油温并附带6个功率负载特征。数据按小时级（ETTh1、ETTh2）和15分钟级（ETTm1）划分子集，训练/验证/测试集比例为12/4/4个月，适用于多变量和单变量时间序列预测任务，由多所高校于2021年发布。