Generalized-VDEP

This is an official implementation of "Generalized-VDEP". (Submitted to IJCSS)

Installation

0. Prerequisites Before running this project, make sure you have the following requirements

• Python 3.7.
• Poetry: This project uses Poetry to manage dependencies. Install Poetry before using this repository.

1. Clone this repository:

$ git clone https://github.com/Rikuhei-ynwa/Generalized-VDEP.git

2. Install dependencies:

$ cd Generalized-VDEP
$ poetry install

3. Install StatsBomb data

$ git clone https://github.com/statsbomb/open-data.git

This open data is often updated, so if you want to reproduce the results of the paper, you should checkout the commit hash of the URL mentioned in the paper.

$ cd ./open-data-master
$ git checkout 533862946a73608c134d18b78226b6371ce7173c

Also, to run the code, you need to rename the directory to open-data-20240702.

Evaluation

You can set the country name in the teamView option. The following command evaluates the performance of the England team in the UEFA EURO 2020 tournament.

$ cd ./Generalized-VDEP
$ poetry run python main_evaluate.py --data statsbomb --game euro2020 --n_nearest 11 --no_games -1 --date_opendata 20240702 --date_experiment ${date_experiment} --model xgboost --test --teamView England

Verification

Attention: This verification process must be performed after the evaluation process.

$ poetry run python main_verify.py --data statsbomb --game euro2020 --date_opendata 20240702 --date_experiment ${date_experiment} --model xgboost --k_fold 5

Results folder

The results of the evaluation and verification are saved in the ./GVDEP_data directory. The resulting directory structure will be as follows:

📁 Generalized-VDEP/
📁 open-data-2024-0702/
📁 GVDEP_data/
└─📁 data-statsbomb/
  └─📁 euro2020/
    └─📁 ${date_experiment}/
      └─(The results of running the code are summarized)

Citation

If our work is useful for your project, please consider citing the paper:

@article{umemoto2022location,
  title={Location analysis of players in UEFA EURO 2020 and 2022 using generalized valuation of defense by estimating probabilities},
  author={Umemoto, Rikuhei and Tsutsui, Kazushi and Fujii, Keisuke},
  journal={arXiv preprint arXiv:2212.00021},
  year={2022}
}

Acknowledgements

We appreciate the following repositories:

• ML-KULeuven/socceraction
• statsbomb/open-data

License

• This software is created under MIT License same as ML-KULeuven/socceraction

Contact

If you have any questions, please contact author:

• Rikuhei Umemoto (umemoto.rikuhei[at]g.sp.m.is.nagoya-u.ac.jp)