# CoupledGNN

**Repository Path**: wu58814/CoupledGNN

## Basic Information

- **Project Name**: CoupledGNN
- **Description**: No description available
- **Primary Language**: Python
- **License**: Not specified
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 1
- **Forks**: 0
- **Created**: 2020-10-20
- **Last Updated**: 2021-12-10

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# CoupledGNN
Adopted for python3.

This repository is an implementation of our proposed CoupledGNN model in the following paper:

```
Qi Cao, Huawei Shen, Jinhua Gao, Bingzheng Wei, Xueqi Cheng. 2020. Popularity Prediction on Social Platforms 
with Coupled Graph Neural Networks. In WSDM'20, February 3-7, 2020, Houston, TX, USA, 9 pages.
```

The CoupledGNN model solves the *network-aware popularity prediction* problem, capturing the cascading effect explicitly by two coupled graph neural networks.

For more details, you can download this paper [Here](https://arxiv.org/abs/1906.09032)

## Requirements

See `requirements.txt`.

~~Python 2.7.5~~

python 3.7(adopted, use `2to3`)

Tensorflow 1.14.0(there are too many bugs while adopted to Tensorflow2+, so i give up)

## Installation

**Install required libraries by `requirements.txt`,  virtual enviroment is recommend.**

Use Conda:

```shell
conda install --file requirements.txt
```

Use Pip:

```shell
pip install -r requirements.txt
```

**Note:**

Tensorflow is need the `Microsoft Visual C++ Redistributable for Visual Studio`, you can download it in https://visualstudio.microsoft.com/zh-hans/downloads/, or click [this(Microsoft Visual C++ Redistributable for Visual Studio 2019 x86)](https://aka.ms/vs/16/release/VC_redist.x64.exe)


## Usage

***Example Usage***

`python -u train.py --dataset=artificial1 --learning_rate=5e-4 --graph_learning_rate=5e-5 --n_layers=3`

Append ` | tee ouput.txt` to above command, for saving the output to file while keeping print to screen(for powershell).

Use `handle_output.py` to plot the result of each training epochs when finished(read data from `output.txt`).

For detailed description of all parameters, you can run

`python -u train.py --help`

## Cite
Please cite our paper if you use this code in your own work:
```
@inproceedings{cao2020coupledgnn,
  title={Popularity Prediction on Social Platforms with Coupled Graph Neural Networks},
  author={Cao, Qi and Shen, Huawei and Gao, Jinhua and Wei, Bingzheng and Cheng, Xueqi},
  booktitle={Proceedings of the 13th ACM International Conference on Web Search and Data Mining},
  series={WSDM'20},
  year={2020},
  location={Houston, TX, USA},
  numpages={9}
}
```