We received four final submissions. You can find the results in the following table:

Rank Teamname Test F1 score
1. voyTECH 0.3433
2. CoolStoryBob 0.2647
3. ItsBoshyTime 0.2593
4. StinkyCheese 0.1422
5. Random Baseline 0.0741

# Next Steps

We provide the test dataset so that you can analyze your model in your paper.

In test_truth.csv, you find multiple additional columns that help you with your analysis:

• channel, user, subscribed as in the training dataset
• channel_activity and user_activity give a class (“low”, “normal”, or “high”) according to the number of messages they write/receive, as explained in the task description.
• messages: number of messages of that user in that channel
• in_train: whether the user is present in the training dataset or not

For your paper submission that is due July 1st (anywhere on earth), please use the LNCS LaTeX template. Your paper should have 6-10 pages and describe your model architecture and training procedure. You can use the test dataset to do further analyses, such as the influence of users being present in the training data or the activity status of channels and users.

In your paper submission, please cite the following paper, which will provide the dataset and task description:

@ARTICLE {kobs2020towards,
author  = "Kobs, Konstantin and Potthast, Martin and Wiegmann, Matti and Zehe, Albin and Stein, Benno and Hotho, Andreas",
title   = "Towards Predicting the Subscription Status of Twitch.tv Users --- ECML-PKDD ChAT Discovery Challenge 2020",
journal = "Proceedings of ECML-PKDD 2020 ChAT Discovery Challenge",
year    = "2020"
}


The title of your paper should contain your team emote name in order to easily connect the corresponding results and paper.

Please send your paper submission as a PDF in time to ecml2020-dc@professor-x.de. We will then review your submission and notify you by July 15th, such that you can register for ECML-PKDD in order to present your work. We will provide further information when it is clear how exactly ECML-PKDD will take place this year.