contrib/papers/DiffNet++

• Contents
• Diffnet++ Description
• Model Architecture
• Dataset
• • Yelp
• Environment Requirements
• Quick Start
• Script Description
• • Script and Sample Code
  • Training Process
  • Evaluation Process
• Model Description
• • Performance
  • • Evaluation Performance
• Description of Random Situation
• ModelZoo Homepage

Contents

[toc]

Diffnet++ Description

Diffnet++ model is a classical model in Social Recommendation area. This is an implementation of Diffnet++ as described in the DiffNet++: A Neural Influence and Interest Diffusion Network for Social Recommendation paper.

Model Architecture

Diffnet++ model fuses both influence diffusion in the social network $G_S$ and interest diffusion in the interest network $G_I$ for social recommendation. The architecture of DiffNet++ contains four main parts: an embedding layer, a fusion layer, the influence and interest diffusion layers, and a rating prediction layer.

Dataset

The Yelp dataset are used for model training and evaluation.

Yelp

Yelp is a well-known on-line location based social network, where users could make friends with others and review restaurants.

Yelp dataset contains 204,448 ratings of 38,342 items by 17,237 users. We transform the original ratings to binary values. If the rating value is larger than 3, we transform it into 1, otherwise it equals 0. Each line of this file has the following format:

user_id item_id label

Environment Requirements

Hardware(GPU）

• Prepare hardware environment GPU processor.

Framework

• MindSpore

For more information, please check the resources below：

• MindSpore Tutorials
• MindSpore Python API

Quick Start

After installing MindSpore via the official website, you can start training and evaluation as follows:

python train.py
python eval.py

Script Description

Script and Sample Code

└── Diffnet_plus_plus
  ├── check_point
    ├── conf
        ├── yelp_Diffnet_plus_plus_ms.ini
    ├── data
    ├── datahelper.py
    ├── DataModule.py
    ├── dataset.py
    ├── eval.py
    ├── train.py
    ├── Metrics.py
    ├── ParserConf.py
    ├── README.md

Training Process

python train.py

After training, you'll get some checkpoint files under the script folder by default. The loss value will be achieved as follows:

Epoch:1
Loss:66502.18
Epoch:2
Loss:66335.62
...

The model checkpoint will be saved in the current directory.

Evaluation Process

Before running the command below, please check the checkpoint path used for evaluation. Please set the checkpoint path to be the absolute full path, e.g., "./check_point/Diffnetplus_220.ckpt".

python eval.py

The above python command will run in the background. You can view the results on the screen. The accuracy of epoch 220 below is the best:

HR: 0.314 NDCG: 0.277

Model Description

Performance

Evaluation Performance

Parameters	GPU
Model Version	Diffnet++
Resource	GPU
Uploaded Date	1/9/2021
MindSpore Version	1.2.1
Dataset	Yelp
batch_size	Full batch
Accuracy	HR: 0.314 NDCG: 0.277

Description of Random Situation

There are three random situations:

• Generating negative samples randomly.

• Initialization of some model weights.

ModelZoo Homepage

Please check the official homepage.