# FilterNet

**Repository Path**: hexin-x/FilterNet

## Basic Information

- **Project Name**: FilterNet
- **Description**: No description available
- **Primary Language**: Unknown
- **License**: Not specified
- **Default Branch**: main
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2025-07-15
- **Last Updated**: 2025-07-15

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# FilterNet (NeurIPS 2024)

The repo is the official implementation for the paper: "[**FilterNet: Harnessing Frequency Filters for Time Series Forecasting**](https://arxiv.org/abs/2411.01623)".


## Getting Started

### 1、Environment Requirements

To get started, ensure you have Conda installed on your system and follow these steps to set up the environment:

```
conda create -n FilterNet python=3.8
conda activate FilterNet
pip install -r requirements.txt
```

### 2、Download Data
All the datasets needed for FilterNet can be obtained from the [Google Drive](https://drive.google.com/drive/folders/1ZOYpTUa82_jCcxIdTmyr0LXQfvaM9vIy) provided in Autoformer. 

### 3、Training Example

For datasets with a small number of variables, such as ETTh, ETTm, and Exchange, we recommend using **PaiFilter** as follows:
```
bash ./scripts/PaiFilter/ETTm1.sh
bash ./scripts/PaiFilter/ETTm2.sh
bash ./scripts/PaiFilter/ETTh2.sh
```

For datasets with a large number of variables such as ECL, Traffic, and weather, it is recommended to use **TexFilter** as follows:
```
bash ./scripts/PaiFilter/ECL.sh
bash ./scripts/PaiFilter/Traffic.sh
bash ./scripts/PaiFilter/Weather.sh
```

## Updates ##
👉 News (2024.12): Another one of our recent works, [**Amplifier: Bringing Attention to Neglected Low-Energy Components in Time Series Forecasting**](https://arxiv.org/abs/2501.17216), has been accepted by AAAI 2025.


## Our Other Work about Learning in the Frequency Domain for Time Series Analysis
🚩 [IJCAI 2024]: [**Deep Frequency Derivative Learning for Non-stationary Time Series Forecasting**](https://arxiv.org/abs/2407.00502)

🚩 [NeurIPS 2023]: [**Frequency-domain MLPs are more effective learners in time series forecasting**](https://proceedings.neurips.cc/paper_files/paper/2023/hash/f1d16af76939f476b5f040fd1398c0a3-Abstract-Conference.html) 

🚩 [NeurIPS 2023]: [**FourierGNN: Rethinking Multivariate Time Series Forecasting from a Pure Graph Perspective**](https://proceedings.neurips.cc/paper_files/paper/2023/hash/dc1e32dd3eb381dbc71482f6a96cbf86-Abstract-Conference.html) 

🚩 [arXiv]: [**A Survey on Deep Learning based Time Series Analysis with Frequency Transformation**](https://arxiv.org/abs/2302.02173)



## Acknowledgement

We appreciate the following GitHub repositories for providing valuable code bases and datasets:

https://github.com/wanghq21/MICN

https://github.com/thuml/TimesNet

https://github.com/aikunyi/FreTS

https://github.com/VEWOXIC/FITS

https://github.com/plumprc/RTSF

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

https://github.com/zhouhaoyi/Informer2020

https://github.com/thuml/Autoformer

https://github.com/ant-research/Pyraformer

https://github.com/MAZiqing/FEDformer

https://github.com/yuqinie98/PatchTST

https://github.com/thuml/iTransformer

https://github.com/thuml/Time-Series-Library