مهيار عبدلي درينسو

The topic and extent of user activity within social networks are influenced by a multitude of factors. As users publish content on these platforms, they contribute to the creation and dissemination of social network events. Examining the factors that impact both the quantity and quality of a user's engagement in social networks not only sheds light on individual characteristics but also unveils behavioral patterns within virtual societies. Moreover, this analysis can assist social decision-makers in making informed choices and provide deeper insights into user interests and preferences. To the best of our knowledge, predicting user participation in social network events while considering constraints such as participation time and order has yet to be explored. This thesis aims to address this gap by introducing a precise definition of social network events. Utilizing implicit feedback and users' participation history, three collaborative filtering-based models are presented to predict user participation in these events, taking into account time constraints. The optimal hyperparameter configurations for each method were meticulously determined, followed by a comprehensive exploration of their significance. Additionally, we conducted a thorough examination of user engagement during various stages of multiple social network events. Notably, our method eva‎luation underscores the strong performance of the presented models, even in the absence of text-based event features.

پالايش مشاركتي , تحليل شبكه اجتماعي , پيش‌بيني كاربر , رويداد شبكه اجتماعي

Collaborative Filtering , Social Network Analysis , User Prediction , Social Network Event

Mahyar Abdoli Darinsu

Dr. Hassan Naderi