See paper in arXiv

We release the PyTorch code of the TTIR model.

• Prerequisites
• Dataset
• Code
• Result

The code is built with following libraries:

• PyTorch 1.0 or higher
• Comet-ml
• PyTorchLightning
• Google Colab

The used dataset is available here.

We develop this project using Google Colab. That's why you must have a Google Account and the dataset in a gDrive folder. Furthermore, you have to change these paths according to the location of the dataset.

train_path = '/content/gdrive/My Drive/Proyecto_RecSys/dataset/train_splits.pkl'
test_path = '/content/gdrive/My Drive/Proyecto_RecSys/dataset/test_splits.pkl'
champion_path = '/content/gdrive/My Drive/Proyecto_RecSys/dataset/champion_types.pkl'

And the comet parameters (api_key, project_name, workspace)

comet_logger = CometLogger(
    experiment_name=conf['exp_name'],
    api_key = 'YOUR_KEY',
    project_name="YOUR_PROJECT_NAME",
    workspace = 'YOUR_WORKSPACE'
)

This work uses the proposed models in Data mining for item recommendation in MOBA games paper as baselines.

This method outperforms the state of the art approaches and explains the result.

If you find this repository useful for your research, please consider citing our paper:

@inproceedings{ttir,
    author = {Villa, Andrés and Araujo, Vladimir and Cattan, Francisca and Parra, Denis},
    title = {Interpretable Contextual Team-aware Item Recommendation]{Interpretable Contextual Team-aware Item Recommendation: Application in Multiplayer Online Battle Arena Games},
    year = {2020},
    isbn = {9781450375832},
    publisher = {Association for Computing Machinery},
    address = {New York, NY, USA},
    url = {https://doi.org/10.1145/3383313.3412211},
    doi = {10.1145/3383313.3412211},
    booktitle = {Proceedings of the 14th ACM Conference on Recommender Systems},
    keywords = {item recommendation, deep learning, MOBA games},
    location = {Virtual Event, Brazil},
    series = {RecSys ’20}
}

For any questions, welcome to create an issue or contact Andrés Villa ([email protected]) - Vladimir Araujo ([email protected]).