Data Challenge: 2022

In this edition of the annual PHME Data Challenge, participants are challenged to demonstrate the application of state-of-the-art algorithms and models to solve a classification problem for a real production line. This year's Challenge is held in collaboration with Bitron Spa, a leading company in the development and manufacturing of mechatronic devices. Thanks to this collaboration, the competition has exclusive access to an extensive dataset of a real industrial production line equipped with automated, integrated and fully connected machines (Industry 4.0).

For more information about the data challenge and to enroll in it, please visit the https://phm-europe.org/data-challenge web page. Here we will post information about the data challenge, the data and basic Jupyter notebooks.

[NEW!!] The test set is now online!

Competition Closed Extension: June 8, 2022, 11:59:59 GMT

A new file called "SPI-ColumnDescription.txt" is now available. It reports the meaning of each column of the SPI dataset. Evaluation process and metrics are now avaiable !! The updated Notebook we will use for evaluation is online as well as the leader board.

For more information about the data challenge, you can issues in this Repository.

Due to the training and validation dataset size, we split the experiments into two different folders. To obtain the password to access the experiment, please register for the data challenge at https://phm-europe.org/data-challenge.

Computing Environment and Leader Board

The participants will have the opportunity to use the online environment for Data Science hosted by Politecnico di Torino, hosted at jupyter.polito.it. Participants in the data challenge are strongly encouraged to use this online environment, which offers persistent storage.

The participants' code will be executed on daily basis and the results shown in the online leader board, indicating the scores of the teams. To have an account or request help, please email datachallenge@phmeurope.org.

Important: When you receive an account, in your home directory in jupyter.polito.it we provide you the data and the example notebook that you have to complete as well as the data. Follow the instructions in the Notebook to provide your solution in the correct format.

The Leader Board will be available soon.

License

This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.

Citation

For citation purposes, please use:

Danilo Giordano, Martino Trevisan. (2022). "PHME Data Challenge". European conference of the prognostics and health management society.