InBox: Recommendation with Knowledge Graph using Interest Box Embedding

This repository contains the code for the paper "InBox: Recommendation with Knowledge Graph using Interest Box Embedding," which has been accepted by VLDB.

Code Organization

The codebase is organized as follows:

• main.py: The main script to run the project.
• model.py: Contains the model definitions.
• dataloader.py: Contains the data loading utilities.
• utils/: Contains utility functions and scripts for data processing and evaluation.

Running the Project

CUDA_VISIBLE_DEVICES=0 python main.py --cuda --dataset=alibaba-fashion -pre -pre_i -train -test

CUDA_VISIBLE_DEVICES=0 python main.py --cuda --dataset=yelp2018 -pre -pre_i -train -test

CUDA_VISIBLE_DEVICES=0 python main.py --cuda --dataset=last-fm -pre -pre_i -train -test

CUDA_VISIBLE_DEVICES=0 python main.py --cuda --dataset=amazon-book -pre -pre_i -train -test -pre_epoch 8

Datasets

The project includes the following datasets:

• Alibaba Fashion
• Amazon Book
• Last.fm
• Yelp 2018

Each dataset directory contains multiple files such as entity_list.txt, item_list.txt, kg_final.txt, etc.

Hardware/Software Requirements

• Python 3.6+
• PyTorch 1.7+
• CUDA 10.1+
• tensorboardX
• scikit-learn
• tqdm

Ensure that you have a compatible GPU and CUDA installed for running the project.