Transformer/self-attention for Multidimensional time series forecasting 使用transformer架构实现多维时间预测
Rerfer to https://github.com/oliverguhr/transformer-time-series-prediction
torch 1.7.0
python 3.7
matplotlib 3.4.1
numpy 1.20.2
After downloading data from https://archive.ics.uci.edu/ml/datasets/ElectricityLoadDiagrams20112014, I convert it into .xlsx file
This is the .xlsx file (source: 'https://pan.baidu.com/s/1gyN37cM_ZI2i7OOaiaQ-_g' password: 4iu4)
You can also download by this link:https://docs.google.com/spreadsheets/d/1P8nIB2Ugd9AMcGlAPu3lD2yh2eAyJVh3/edit?usp=drive_link&ouid=102877252444231070619&rtpof=true&sd=true or https://drive.google.com/file/d/13FyJqP_MVVHzqQ3G0egpglelO0G-dsEa/view?usp=drive_link
And in my code, I select 10 dims time series for modeling
https://zhuanlan.zhihu.com/p/390193679