Simperl, ElenaShi, MiaojingAlghamdi, Kholoud2024-08-182024-08-182024-08-01https://hdl.handle.net/20.500.14154/72885Wikidata is an open knowledge graph built by a global community of volunteers. As it advances in scale, it faces substantial challenges around editor engagement. These challenges are in terms of both attracting new editors and retaining existing ones. Experience from other online communities and peer-production systems, including Wikipedia, suggests that personalised recommendations could help contributors keep up with the sheer amount of work, especially newcomers, who are sometimes unsure about how to contribute best to an ongoing effort. For these reasons, in this thesis, we propose personalised task recommendation systems for Wikidata editors. To the best of our knowledge, this is the first work of its kind for Wikidata. Therefore, our focus is to create a foundation for understanding how Wikidata items can be recommended to the editors. To achieve this, our research comprises three key phases: understanding editors’ practices and translating them into system design requirements, developing recommendation models based on these requirements, and evaluating the models with Wikidata editors. This thesis makes four main contributions: (i) it provides insights into the needs, preferences, and practices of Wikdiata editors, aiding in the formulation of design requirements for our task recommendation systems; (ii) it develops three different recommendation models that can recommend items and topics to the Wikidata editors, considering past editing activities, the features of the items and the long and short editor interests; (iii) it creates two benchmark datasets that can facilitate further research in the area of Wikidata recommendations; (iiii) it provides an evaluation which demonstrates that the proposed recommender models produce more relevant recommendations for the editors than other tools in Wikidata. Our findings show the feasibility of implementing personalised recommendations in Wikidata, and they pave the way for further research to improve the recommendations and leverage them to enhance the quality of the knowledge graph.148enWikidataThesis