期刊
出版社
ASSOC COMPUTING MACHINERY
DOI: 10.1145/3219819.3220043
关键词
Word embeddings; Dynamic model; Profiling
类别
资金
- King Abdullah University of Science and Technology (KAUST), Saudi Arabia
In this paper, we study the problem of dynamic user profiling in Twitter. We address the problem by proposing a dynamic user and word embedding model (DUWE) and a streaming keyword diversification model (SKDM). DUWE dynamically tracks the semantic representations of users and words over time and models their embeddings in the same space so that their similarities can be effectively measured. We utilize Bamler and Mandt's skip-gram Filtering algorithm [4] for our inference, which works with a convex objective function that ensures the robustness of the learnt embeddings. SKDM aims at retrieving top-K relevant and diversified keywords to profile users' dynamic interests. Experiments on a Twitter dataset demonstrate that our proposed embedding algorithms outperform state-of-the-art non-dynamic and dynamic embedding and topic models.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据