AnaNET: Anatomical Network for Aggregated Time Series Forecasting in Multi-Layered Architecture

This repo is the official Pytorch implementation of AnaNET.

Getting Started

Environment Requirements

First, please make sure you have installed Conda. Then, our environment can be installed by:

conda env create -f environment.yml
conda activate AnaNET

Data Preparation

We provide the AggTS dataset used in the experiments in the ./dataset directory.

Training Example

• In scripts/ , we offer scripts for training datasets related to DCP and ETT.

For example:

To train the AnaNET on MD dataset, you can use the scipt scripts/run_DCP.sh:

sh scripts/run_DCP.sh

Acknowledgment

We appreciate the following GitHub repos a lot for their valuable code base or datasets:

https://github.com/thuml/Autoformer

https://github.com/cure-lab/LTSF-Linear

https://github.com/yuqinie98/PatchTST

https://github.com/EdgeBigBang/EasyTS

https://github.com/MAZiqing/FEDformer

https://github.com/alipay/Pyraformer