Skip to main content
  • Expression of Interest

    DonanimHaber - Social Network Platform in Turkey

    Our study area generally deals with NLP problems on deep learning-based artificial neural networks such as toxic content detection (together with sensitive content classification on rich media data), text similarity matching and anomaly detection. Furthermore, our team developed and integrated a hybrid recommender engine to suggest content to our users, a semantic search engine used to find relevant threads in platform, and a smart advertisement project to display ads relevant with the contents in platform.

    DonanimHaber is one of the largest social media platforms and a technology news website in Turkey. We're serving more than 125+ million pages to 11 million users monthly (via google analytics)

    The platform has millions of text data (and also images) and that is increasing day by day. A total of 2,224,319 members registered and sent 150+ million messages in 1024 forums (categories). There are currently 11,056.395 threads in total.

    The platform provides a rich source of Turkish natural language, which is an under-represented language used by almost 300 millions of people worldwide. With this feature, our platform provides various research opportunities to social science and computational linguistics researchers.

    With its data science team, Donanim Haber has been developing government-supported or self-funded projects under the title of R&D center in Turkey since 2017. Since then, the team collaborated with academia to develop numerous funded data science projects mostly focusing on NLP and social science related topics. Some of these projects include detection of sensitive contents, AI based advertising systems, AI based search engines and hybrid recommender systems.

    With this current CHISTERA call, we aim to discover opportunities to prevent misbehavior and provide friendly and healthy social network experience for our large user base. In addition, we hope to find an opportunity to share our findings to academia via publications.