# IR2-Net

**Repository Path**: Hanwen-Hu/IR2-Net

## Basic Information

- **Project Name**: IR2-Net
- **Description**: Iterative Time Series Imputation by Maintaining Dependency Consistency (ACM TKDD 2024)
- **Primary Language**: Python
- **License**: Apache-2.0
- **Default Branch**: main
- **Homepage**: https://dl.acm.org/doi/10.1145/3698107
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2024-08-27
- **Last Updated**: 2024-12-07

## Categories & Tags

**Categories**: Uncategorized

**Tags**: Data-Mining, Time-Series-Analysis

## README

# Iterative Time Series Imputation by Maintaining Dependency Consistency

## Datasets
This repository contains the code of `IR^2-Net` with four datasets. Air Quality, Human Activity, Traffic Speed are completely available, but Solar Energy dataset should be firstly unzipped.

## Run
If you want to reproduce the experiment, please run "main.py" with Pycharm or with command
```angular2html
python3 main.py
```
The default dataset is Air Quality with 20% missing data. For the other cases, please add arguments such as
```angular2html
python3 main.py -model GAN -dataset traffic -r_miss 0.4 -cuda_id 0 -use_irm 1 -iter_time 2
```
The argument `use_irm` determines whether to use the incomplete representation mechanism, while `iter_time` represents the time of reconstruction, which are two main contributions of this work.

## Contact
If you have any questions or suggestions for our paper or codes, please contact us. Email: hanwen_hu@sjtu.edu.cn.